Total Phenolic and Total Flavonoid Content, Individual Phenolic Compounds and Antioxidant Activity in Sweet Rowanberry Cultivars

Sweet rowanberry and its cultivars represent a less-known fruit species with significant antioxidant activity, mostly promoted by polyphenolic compounds. This paper examined seven Sorbus cultivars and evaluated their total polyphenolic and flavonoid content, as well as the content of individual polyphenolic compounds from the group of phenolic acids and flavonoids. It also determined their antioxidant activity using DPPH, ACW and ACL. Furthermore, to reflect the distribution of the contribution to antioxidant activity, correlations between antioxidant activity and the contents of ascorbic acid, vitamin E and individual phenolic compounds were established. The highest total phenolic content of 8307.4 mg kg−1 was determined in ‘Granatina’, with the main contribution of phenolic acid content of 7001.7 mg kg−1 and a significantly lower total flavonoid content of 1304.6 mg kg−1. Flavanols represented the most abundant group of flavonoids, with catechin being the second most frequent flavanol with the highest content of 633.67 mg kg−1 in ‘Granatina’. Flavonols were represented by rutin and quercetin. ‘Businka’ displayed a significant vitamin E content of 4.77 mg kg−1, and ‘Alaja Krupnaja’ had the highest vitamin C level of 7.89 g kg−1. These results emphasize their potential health and nutritional benefits and, thus, their promising and valuable role in the food processing industry.


Introduction
The genus Sorbus includes five diploid species in Europe, namely Sorbus aria, Sorbus aucuparia, Sorbus torminalis, Sorbus chamaemespilus and Sorbus domestica [1]. Such diversity stems from interspecific hybridizations [2]. Sweet rowanberry cultivars have been evolved from rowanberries (Sorbus aucuparia L.) and hybrids of rowanberry with Malus, Pyrus, Aronia or Mespilus [3]. Even though the fruits of wild rowanberry excel with their high nutritional values, they have not been recommended for direct consumption due to their specific astringent taste caused mainly by tannins [4]. The first known attempts to select new sweet rowanberry cultivars ('Rossica' and 'Rosina') without astringency and bitterness date back to the 19th century in the Sudeten Mountains, a current territory of the Czech Republic. A breeding program of sweet rowanberries suitable and adaptable for the northern conditions, specifically winter hardiness, was commenced by Michrin in Russia at the beginning of the 20th century. This program introduced new hybrids of rowanberry with the Aronia, Malus, Mespilus and Pyrus species [3]. The most popular cultivars from Russia are 'Burka', 'Likjornaja', 'Dessertnaja', 'Granatnaja', 'Rubinovaja' and 'Titan' [5]. In comparison with wild species, 'Granatnaja' and 'Alaja Krupnaja' are characterized by

Fruit Samples
The experiment samples included the fruit of the following sweet rowanberry cultivars: 'Alaja Krupnaja', 'Granatnaja', 'Granatina', 'Businka', 'Koncentra' and 'Discolor'. Their origin, breeding background and description are provided in Table 1. When compared with wild species, the selected Sorbus cultivars are characterized by an enhanced taste without any astringency or bitterness and by the large size of their fruits. Berries of the tested cultivars were collected from a minimum of five plants in the amount of 500 g per cultivar in a fully ripe state. First, fresh berries were homogenized using a blender (Bosch MSM67170, Bosh GmbH, Stuttgart, Germany) and deep-frozen in an ultra-low temperature freezer (ULUF P610 GG-Arctiko, Esbjerk, Denmark) at −80 • C for at least 24 h. Afterwards, they were lyophilized by Alpha 1-4 LSC (Christ Gefriertocknungsanlagen Gmb, Osterode am Harz, Germany) at −55 • C and 0.120 Mbar for 48 h. Lyophilized samples were homogenized and maintained in polyethylene bags equipped with a zip at −20 • C until they were subjected to the analysis. Round-shaped fruit, dark red or violet, without typical aroma, 2 g, suitable for processing, September-October 'Koncentra' Germany Round-shaped fruit, orange, with a sour taste due to the significant content of vitamin C, September-October

Experimental Area
The experimental area owned by the Mendel University in Brno is in the cadastral area of Žabčice at a latitude of 185 m a.s.l. with the GPS coordinates 49.011598 N and 16.602572 E. It is characterized by typical continental climatic conditions, with a long-term average annual temperature of 9.2 • C and a precipitation of 519.0 mm. In 2014, the average annual temperature in Žabčice was 11.2 • C, and the precipitation was 576.7 mm. During the ripening period from May to July 2014, the average temperature was 18.3 • C, and the precipitation value was 191.2 mm. Table 2 displays long-term average values and annual temperature and precipitation values for 2014 in this experimental area. Fruits were harvested in 2014. This year was characterized by above-average rainfall levels and higher temperatures; specifically, the average temperature exceeded the longterm average by 2 • C, and the precipitation level was 97 mm higher than the long-term average amount. The distribution of rainfall was irregular: from January to April, precipitation was below average; during July, August and September, it was well above average and accounted for more than half of the annual total amount [16].

Chemicals and Reagents
The chemicals ethanol, methanol and acetic acid were obtained from Penta (Prague, Czech Republic) and methanol-HPLC from LabScan (Sowińskiego, Poland). Phenolic compound standards of HPLC grade were acquired from Sigma Aldrich (St. Louis, MO, USA). Standards of ascorbic acid and D-α-tocopherol succinate were acquired from Ac-cuStandard (New Haven, CT, USA). Further used chemicals were of analytical grade and purchased from Sigma Aldrich (St. Louis, MO, USA).

Extraction Methods
The extractions were conducted according to the protocols published by Orsavová et al. (2019) and Sytařová et al. (2020) [17,18]. Lyophilized fruit samples of 0.5 g were extracted in 10 mL of the extraction solution of water and methanol (70/30, v/v) to determine total phenolic content, total flavonoid content and antioxidant activity (by DPPH method). To prepare the extracts for HPLC assay, the same amount of 0.5 g of lyophilized fruit samples was extracted in a mixture of redistilled water/methanol/acetic acid (69/30/1, v/v/v) using screw-cap test tubes in a water bath (Memmert GmbH + Co.KG, Schwabach, Germany) and shook at 50 • C for 60 min. The extracts were then centrifuged at 2430× g for 15 min (Velocity 13µ, Dynamica Scientific Ltd., Milton Keynes, UK) at room temperature. Further, the extracts for vitamin C determination were prepared from 0.5 g of lyophilized fruits using 2.5 mL of a mobile phase (methanol/H 3 PO 4 /redistilled water (99/0.5/0.5, v/v/v)) in screw-cap test tubes in a shaker LT 2 (Kavalier, Sázava, Czech Republic) for 10 min in the dark. The extracts were then added into 10 mL volumetric flasks and filled with a mobile phase. Similarly, the extracts for vitamin E determination were prepared from 1.0 g of lyophilized fruit samples using 2.5 mL of methanol in screw-cap test tubes in an ultrasonic bath PS 04,000 A (Notus-Powersonic, Vráble, Slovakia) at 40 • C for 60 min. Afterwards, the extracts were added into 10 mL volumetric flasks and filled with methanol. Redistilled water was obtained by PURELAB Classic (ELGA, Lane End Business Park, High Wycombe, UK). The extraction for the analysis of total anthocyanin content was performed according to the protocol reported by Orsavová et al. (2019) [17]. Briefly, lyophilized fruit samples of 1.5 g were extracted in 5 mL of a mixture of methanol/water/acetic acid (70/29/1, v/v/v) in screw-cap test tubes in a shaking water bath (Memmert GmbH + Co.KG, Schwabach, Germany) at 50 • C for 60 min and subsequently in an ultrasonic bath at 40 • C for 60 min. The extracts were centrifuged at 3280× g for 15 min (Velocity 13µ, Dynamica Scientific Ltd., Milton Keynes, UK) at room temperature. Prior to the analyses, all extracts and supernatants were filtered using nylon microfilters (SYRINGE, Cronus Syringe Filter, Nylon 13 mm × 0.45 µm, Labicom, Olomouc, Czech Republic).

Analysis of Total Phenolic (TPC) and Total Flavonoid (TFC) Content
According to protocols by Orsavová et al. (2019), total phenolic content (TPC) was established employing the Folin-Ciocalteu method, and total flavonoid content (TFC) applying NaNO 2 , AlCl 3 ·6H 2 O and NaOH using a UV/VIS spectrometer Lambda 25 (PerkinElmer, Waltham, MA, USA) [17]. The results were expressed as grams of gallic acid equivalent per kg (g GA kg −1 DW) for TPC and as grams of rutin equivalent per kg (g RE kg −1 DW) for TFC.

Analysis of Vitamins C and E
The contents of vitamins C and E were recorded following the protocols reported by Orsavová et al. (2019) and Sytařová et al. (2020). The analysis employed an HPLC system, UltiMate ® 3000 (Dionex, Sunnyvale, CA, USA), equipped with a diode-array detector [17,18]. For the vitamin C analysis, the reverse-phase column Acclaim 120 C8 (Dionex, MA, USA) with dimensions of 150 × 2.1 mm and a particle size of 5 µm was applied. The mixture of methanol/H 3 PO 4 /r-H 2 O was used in a ratio of 99:0.5:0.5 (v/v/v) as a mobile phase in an isocratic mode, and the flow rate was set to 0.8 mL min −1 with an injection volume of 20 µL and column temperature of 25 • C maintained throughout the 10 min analysis. Chromatograms were registered at 275 nm. For the analysis of vitamin E, the column Kinetex C-18 (Phenomenex, Torrance, CA, USA) with dimensions of 150 × 4.6 mm and a particle size of 2.6 µm was applied. A mixture of methanol (HPLC)/r-H2O was used in a ratio of 95:5 (v/v) as a mobile phase in an isocratic mode, and the flow rate was set to 1 mL min −1 with an injection volume of 20 µL and column temperature of 30 • C maintained throughout the 20 min analysis. Chromatograms were monitored at 230 nm. Then, vitamin C and E contents were calculated using the calibration curves employing ascorbic acid and D-alpha-tocopherol succinate as standards. The contents of vitamin C were expressed in g kg −1 DW and those of vitamin E in mg kg −1 DW.

Determination of Phenolic Compounds Using HPLC
Individual phenolic compounds were determined using an HPLC device, UltiMate ® 3000 (Dionex, Sunnyvale, CA, USA), equipped with a diode-array detector and Kinetex column C-18 (Phenomenex, Torrance, CA, USA), following the procedure described by Orsavová et al. (2019) [17]. Solvent A was a mixture of water/acetic acid prepared in a ratio of 99:1 (v/v) and solvent B consisted of water/acetonitrile/acetic acid in a ratio of 67:32:   [17] employing DPPH (2,2-diphenyl-1-picrylhydrazyl; Sigma Aldrich, MO, USA) and PCL. During DPPH analysis, the absorbance was recorded at 515 nm by Lambda 25 (PerkinElmer, Waltham, MA, USA). Trolox, obtained at Sigma Aldrich, MO, USA, was used as a standard. The results were expressed as grams of Trolox equivalent per kg (g Trolox kg −1 DW). For the PCL analysis, official ACW and ACL protocols were applied using ACW and ACL kits (Analytik Jena AG, Jena, Germany) for water-soluble and lipid-soluble compounds, respectively, in PHOTOCHEM (Analytik Jena AG, Jena, Germany). ACW and ACL were quantified by applying calibration curves. Ascorbic acid was used as a standard for ACW and Trolox for ACL, with the results expressed as grams of ascorbic acid equivalent per kg (g AA kg −1 DW) or Trolox equivalent (g Trolox kg −1 DW).

Statistical Analysis
All experiments were repeated three times, and their results were expressed as means and standard deviations. SPSS 12.0 (SPSS Inc., Chicago, IL, USA) was applied to confirm significant differences between the examined values. The Shapiro-Wilk test was used to test normal data distribution. Then, for normally distributed data, one-way analysis of variance (Anova, Tukey's test) was employed with the level of significance set to p < 0.05. For abnormal data distribution, a non-parametric Kruskal-Wallis test was used with the same significance level. Pearson correlation coefficients (R) were calculated using Microsoft Office Excel 2013 (Redmond, WA, USA), and Evans' classification was applied to assess the strength of correlations [20].

Determination of Total Polyphenolic (TPC), Total Flavonoid (TFC) and Anthocyanin Content (AC)
The values of total polyphenolic (TPC), total flavonoid (TFC) and anthocyanin content (AC) in the examined samples of sweet rowanberry significantly differed between the cultivars, as can be seen in Table 3.  [21]. Lower TPC values were determined in 'Alaja Krupnaja' (6.46 ± 0 g GA kg −1 ) and 'Bussinka' (2.88 ± 0 g GA kg −1 ) in comparison with the results of this study, which determined TPC in the amounts of 9.00 ± 0.02 g GA kg −1 in 'Titan'. This is in alignment with the data recorded by Hukkinen et al. (2006), who determined TPC in the amounts of 8.08 ± 1.9 g GA kg −1 [3].
On the other hand, significantly lower TPC values were reported by Jabłońska-Ryś et al. (2009) when they studied samples of sweet rowanberry from Poland [22]. Similarly, in the cultivar 'Granatnaja' from the Czech Republic, lower TPC values of 3.65 g GA kg −1 FW were recorded in 2011 and 2012 [12].
A comparable TPC content of 8.19 g GA kg −1 FW was established in 'Granatnaja' from the Czech Republic in 2011 and 2012. The lower values of 8.11 g GA kg −1 FW in 'Granatina' and 6.28 g GA kg −1 FW in 'Titan' proved the generally variable composition of sweet rowanberry fruits [6]. The total phenolic content in Sorbus has generally shown a strong dependence on the maturity stage of the berries, while the recovery of phenolics has been influenced by the extraction solvent [23].
This study evaluated the significant differences in total flavonoid content (TFC) among all assayed cultivars. The lowest content of 15.04 g RU kg The results of this experiment have clarified statistically significant differences in anthocyanin content (AC). Light-colored 'Alaja Krupnaja' displayed a very low anthocyanin content at 1.19 mg COG 100 g −1 , together with 'Koncentra' and 'Discolor', which had 5.36 mg COG 100 g −1 and 5.50 mg COG 100 g −1 , respectively. In contrast, the highest values of 51.38 mg COG 100 g −1 and 50.20 mg COG 100 g −1 were recorded in 'Granatnaja' and 'Titan', respectively.

Total Contents of Phenolic Compounds by RP-HPLC
The total phenolic (TPC), total flavonoid (TFC) and total phenolic acid contents (TPA) are listed in Figure  According to Kampuss et al. (2009), the total phenolic content ranged between 162 mg 100 g −1 of fresh weight in 'Krasnaja Krupna' and 485 mg 100 g −1 in 'Likernaja' in samples from Latvia [27].
In the studies by Mattila et al. (2006), the average TPC of 24 analyzed sweet rowanberry fruit samples reached the highest value of 103 mg 100 g −1 FW [11]. In all the assayed cultivars, TPA ranged from 524.7 mg kg −1 in 'Discolor' to 7001.7 mg kg −1 in 'Granatina', which significantly contributed to TPC. The study by Mattila et al. (2006) comparing the phenolic profiles of berries established an average phenolic acid content of 75 mg 100 g −1 . In this study, the total flavonoid content ranged from 332 mg kg −1 in 'Discolor' to 1304.6 mg kg −1 in 'Granatina' [11].

Individual Phenolic Compounds by RP-HPLC
Phenolic acids have been observed as the most abundant phenolic substances in Sorbus. On the other hand, flavonoids have shown lower contents [28].
As the major flavonoids, quercetin, isoquercetin, kaempferol, rutin, hyperoside and isorhamnetin were reported in the examined samples of fruits, leaves and inflorescences of the selected Sorbus cultivars [4].
The contents of individual flavonoids, stilbenes and resveratrol, together with the sum of total flavonols (TFLOC) and flavanols (TFLAC) are provided in Table 4 and Figure  2.
In the studies by Mattila et al. (2006), the average TPC of 24 analyzed sweet rowanberry fruit samples reached the highest value of 103 mg 100 g −1 FW [11]. In all the assayed cultivars, TPA ranged from 524.7 mg kg −1 in 'Discolor' to 7001.7 mg kg −1 in 'Granatina', which significantly contributed to TPC. The study by Mattila et al. (2006) comparing the phenolic profiles of berries established an average phenolic acid content of 75 mg 100 g −1 . In this study, the total flavonoid content ranged from 332 mg kg −1 in 'Discolor' to 1304.6 mg kg −1 in 'Granatina' [11].

Individual Phenolic Compounds by RP-HPLC
Phenolic acids have been observed as the most abundant phenolic substances in Sorbus. On the other hand, flavonoids have shown lower contents [28].
As the major flavonoids, quercetin, isoquercetin, kaempferol, rutin, hyperoside and isorhamnetin were reported in the examined samples of fruits, leaves and inflorescences of the selected Sorbus cultivars [4].
The total flavanol content represented a significant proportion of flavonoids and was present in amounts ranging from 298.8 mg kg −1 in 'Discolor' to 1668.6 mg kg −1 in 'Koncentra'.
Epigallocatechin (EGC) was the predominant flavanol in four of the analyzed cultivars, with the highest amount of 1167.5 mg kg −1 in 'Koncentra'; the rest of the samples showed amounts ranging from 244.3 mg kg −1 in 'Discolor' to 625.4 mg kg −1 in 'Businka'. The catechin (CA) content ranged from 23.4 mg kg −1 in 'Discolor' to 633.6 mg kg −1 in 'Granatina', while epicatechin (EC) was present in smaller amounts (from 3.3 mg kg −1 in 'Koncentra' to 31.1 mg kg −1 in 'Discolor'). A significantly lower EC concentration of 0.38 mg kg −1 was identified in the samples of Sorbus umbellata from Turkey [24], in contrast to a very high amount of 862.50 mg kg −1 recorded in the samples from the Czech Republic in 2008 and 2010 [31].
Stilbene resveratrol (RES) was established only in very low amounts. The highest content of 3.3 mg kg −1 was determined in 'Alaja Krupnaja' and the lowest of 0.5 mg kg −1 in 'Titan'. It was not identified in the cultivar 'Discolor', similarly to the Sorbus umbellata samples from Turkey [24].
A significant diversity in the profile of flavonols has been observed depending on the particular cultivar (Table 4, Figure 2). All rowanberry fruit powder samples showed the presence of rutin, hyperoside and isoquercitrin [26].
Quercetin was detected only in the cultivar of 'Alaja Krupnaja' in the amount of 2.4 mg kg −1 which is in contrast with its amount in the 26 cultivars from Serbia and Montenegro ranging from 2.8 mg kg −1 to 83.5 mg kg −1 [10] and also its content of 1.30 mg kg −1 found in Sorbus umbellata from Turkey [24].  [24].
The quercetin content in the fruits of S. aucuparia, S. intermedia and S. aria was 0.51, 0.31 and 0.09 mg g −1 , respectively. Kaempferol was quantified in the fruits, leaves and inflorescences of the same Sorbus species, with its highest content recorded in S. aucuparia [30]. nd-not detected. The results are expressed as arithmetic means ± SD (n = 6). The values in a row with different superscripts indicate a statistically significant difference at the significance level of p < 0.05.
When compared with the results of this study, a significantly higher value of 1290 mg kg −1 of total flavonols was found in 'Granatnaja' from Finland by Kylli et al. (2010).
The total flavanol content represented a significant proportion of flavonoids and was present in amounts ranging from 298.8 mg kg −1 in 'Discolor' to 1668.6 mg kg −1 in 'Koncentra'.
Epigallocatechin (EGC) was the predominant flavanol in four of the analyzed cultivars, with the highest amount of 1167.5 mg kg −1 in 'Koncentra'; the rest of the samples showed amounts ranging from 244.3 mg kg −1 in 'Discolor' to 625.4 mg kg −1 in 'Businka'. The catechin (CA) content ranged from 23.4 mg kg −1 in 'Discolor' to 633.6 mg kg −1 in 'Granatina', while epicatechin (EC) was present in smaller amounts (from 3.3 mg kg −1 in 'Koncentra' to 31.1 mg kg −1 in 'Discolor'). A significantly lower EC concentration of 0.38 mg kg −1 was identified in the samples of Sorbus umbellata from Turkey [24], in contrast to a very high amount of 862.50 mg kg −1 recorded in the samples from the Czech Republic in 2008 and 2010 [31].
Stilbene resveratrol (RES) was established only in very low amounts. The highest content of 3.3 mg kg −1 was determined in 'Alaja Krupnaja' and the lowest of 0.5 mg kg −1 in 'Titan'. It was not identified in the cultivar 'Discolor', similarly to the Sorbus umbellata samples from Turkey [24].
A significant diversity in the profile of flavonols has been observed depending on the particular cultivar (Table 4, Figure 2). All rowanberry fruit powder samples showed the presence of rutin, hyperoside and isoquercitrin [26].

Determination of Phenolic Acids
The greatest amounts of phenolic acids were found in rowanberries by Mattila et al. (2006), in contrast with their amounts in the samples of chokeberry, saskatoon berry, blueberry, raspberry, bilberry, cloudberry, rosehip, lingonberry, black currant and bog whortleberry [11]. Mrkonjic et al. (2019) used LC-MS/MS to determine 15 phenolic acids, with the predominance of chlorogenic acid in the fruits of S. domestica [32].
Chlorogenic (3-O-caffeoylquinic acid, 3-CQA) and neochlorogenic (5-O-caffeoylquinic acid, 5-CQA) acids represent the most abundant phenolic acids [10], accounting for 56-80% of the total phenolic content in Sorbus berries [8]. Cinnamic, vanillic, p-coumaric and benzoic acids were detected only in trace amounts in the fruits of S. aucuparia [33] and S. domestica [34]; p-coumaric acid was determined in S. discolor fruits as well [35]. Raudonis et al. (2014) established a significant variability in the contents of phenolic acid and flavonoid substances and the levels of antioxidant activity in the fruits of the Sorbus species, which is in accordance with the results of this study [29].
The derivatives of benzoic acids (DBA) were present in the samples of the selected Sorbus cultivars in smaller amounts, ranging from 56.8 mg kg −1 in 'Discolor' to 154.7 mg kg −1 in 'Granatina'. The higher amount of 151.4 mg kg −1 was also determined in 'Alaja Krupnaja'. This study identified significant differences in the presence of individual phenolic acids belonging to the group of derivatives of benzoic acid. For example, gallic acid (GA) was recorded ranging from 1.6 mg kg −1 in 'Discolor' and 'Titan' to 16.7 mg kg −1 in 'Koncentra'. Gallic acid was not detected in the samples of Sorbus umbellata from Turkey [24]. As the prevailing acid from the group of benzoic acid derivatives, vanillic acid (VA) was measured in the amount of 37.  [24].
When compared with benzoic acid derivatives, the derivatives of cinnamic acid (DCA) were present in higher amounts, ranging from 467.9 mg kg −1 in 'Discolor' to 6847.1 mg kg −1 in 'Granatina'. Considerable amounts were also observed in the cultivars 'Granatnaja' and 'Titan', in amounts of 6292.4 mg kg −1 and 6265.1 mg kg −1 , respectively.
Chlorogenic (3-O-caffeoylquinic acid, 3-CQA) (CHA) and neochlorogenic (5-Ocaffeoylquinic acid, 5-CQA) (NCHA) acids were the main phenolic acids in sweet rowanberry, which is in alignment with the results of this study [10,21]. Both of these acids are considered markers of the phytochemical and antioxidant profiles of Sorbus fruits and were present in all the samples of the analyzed Sorbus cultivars [29].
Chlorogenic acid was determined to be the most abundant phenolic compound in S. aucuparia fruits [36] with its content of 200 mg 100 g −1 [3] which is in accordance with this study. The content of CHA ranged from 189.7 mg kg −1 in 'Discolor' to 2375.2 mg kg −1 in 'Titan'. A high amount of CHA was also detected in 'Koncentra' and 'Granatina', with values of 2277.4 mg kg −1 and 2271.9 mg kg −1 , respectively.
Coffeic acid (CA) dominated in 'Alaja Krupnaja' with a value of 1803.6 mg kg −1 ; in the rest of the samples, it ranged from 51.8 mg kg −1 in 'Titan' to 667.4 mg kg −1 in 'Granatnaja'. Kivrak et al. (2014) recorded only an insignificant content of coffeic acid with a value of 3.03 mg kg −1 in the samples from Turkey [24].
The highest content of FEA in the amount of 115.8 mg kg −1 was present in 'Alaja Krupnaja'. Additionally, it ranged from 4.3 mg kg −1 in 'Titan' to 9.1 mg kg −1 in 'Granatina'. However, it was not detected in the cultivar 'Discolor'. Such low values are in accordance with the amount of 7.67 mg kg −1 detected in the samples of Sorbus umbellata from Turkey [24]. P-coumaric acid (PC) was present only in low amounts, ranging from 2.4 mg kg −1 in 'Alaja Krupnaja' to 13.6 mg kg −1 in 'Titan'. It was not detected in Sorbus umbellata fruits [24]. Syringic acid (SA) ranged from 4.1 mg kg −1 in 'Titan' to 61.8 mg kg −1 in 'Granatina'. Hydroxycinnanic acid (HCA) was not detected, and t-cinnamic acid (TCA) was recorded only in a trace amount of 0.9 mg kg −1 in 'Alaja Krupnaja'.
This study establishes the correlations between total phenolic content (TPC), total flavonoid content (TFC), total anthocyanin content (AC) and individual phenolic compounds, with the values of these correlation coefficients displayed in Tables 6 and 7. Table 6. Correlation coefficients (R) between total phenolic content (TPC), total flavonoid content (TFC), total anthocyanin content (AC) and individual phenolic compounds in the selected Sorbus cultivars.  As Table 6 shows, the relation between TPC and TFC can be evaluated as a very weak positive linear correlation (R = 0.0293).

TPC TFC AC
It has been found that the rise in anthocyanin content in the cultivated species does not correspond with enhanced antioxidant activity [8]. Similarly, this study has shown a linear correlation between TPC and AC of R = −0.4034.
Regarding An indirect linear correlation between AC and RU with a negative value of R = −0.3693 was identified; furthermore, a very strong linear correlation between AC and catechins was established (R = 0.6988). The correlations between AC and catechins were very weak (R = 0.0217 for EC and R = −0.1275 for EGC). The correlation between flavanols (FLAVAN) and AC can be evaluated as a linear direct correlation with R = 0.2776.
As is evident from Table 7, the phenolic acids from the DBA group influenced TPC more than the acids from the DCA group. A direct linear correlation between TPC and DBA was established (R = 0.9990). On the other hand, DCA displayed only a weak indirect linear correlation of R = −0.1457. Regarding DBA, the strongest correlations were identified for HB (R = 0.5977) and SI (R = 0.5514); VA, PC and PCEE displayed weaker direct correlations of R = 0.3074, R = 0.1356 and R = 0.2504, respectively. Finally, GA and EL showed very weak indirect correlations of R = −0.1277 and R = −0.0218, respectively.
In the DCA group, PCA displayed the strongest indirect correlation of R = −0.5892. Furthermore, CHA and NCHA showed weaker negative correlations (R = −0.2836 and R = −0.2543, respectively). Positive linear correlations were identified between TPC and CA (R = 0.4222), between TPC and FEA (R = 0.4735) and, lastly, between TPC and SA (R = 0.3555).

Determination of Vitamin C and E
The values of vitamins C and E in the lyophilized samples of the selected Sorbus cultivars are provided in Table 8. The results are expressed as arithmetic means ± SD (n = 6). The values in a row with different superscripts indicate a statistically significant difference at the significance level of p < 0.05.
The content of ascorbic acid in the samples of sweet rowanberry cultivars was monitored in the amount of 12-21 mg 100 g −1 in 'Granatnaya' and 86 mg 100 g −1 in 'Zholtaya' [38,39]. The study by Ozolina and Kampuse (2019) showed the content of vitamin C in the sweet rowanberry juice residues in the amounts of 60.56 ± 5.33 mg 100g −1 [40].
As can be seen from the results in Table 8, vitamin C contents showed statistical differences between the cultivars, ranging between 4.87 g kg −1 in 'Titan' and 7.9 g kg −1 in 'Alaja Krupnaja'. The average reported vitamin C content in sweet rowanberries reached a value of 4.85 g kg −1 FW, which is comparable with its content in 'Titan' [40].
Generally, vitamin E has been recorded in lower amounts than vitamin C. The lowest vitamin E content of 1.42 mg kg −1 was detected in the cultivar 'Alaja Krupnaja'. Otherwise, its content ranged from 3.96 mg kg −1 in 'Titan' to 4.77 mg kg −1 in 'Businka'.

Antioxidant Activity by DPPH, ACW and ACL
The values of DPPH, ACW and ACL in the lyophilized samples of the selected Sorbus cultivars are presented in Table 9. The results are expressed as arithmetic means ± SD (n = 6). The values in a row with different superscripts indicate a statistically significant difference at the significance level of p < 0.05.
As Table 9 shows, the antioxidant activity determined by DPPH displayed lower values in comparison with the results obtained by ACW and ACL. What is more, statistically significant differences between the cultivars were established in nearly all samples. DPPH values ranged from 3.32 g Trolox kg −1 in 'Discolor' to 16.16 g Trolox kg −1 in 'Businka'. The DPPH radical-scavenging activity of the evaluated samples from Latvia reported by Kampuss et al. (2008) ranged from 2.5 g to 11.2 g per g of DPPH radical determined in 'Alaja Krupnaja' and 'Likernaja', respectively [27].
Compared with the other methods, the ACW method provided the most significant values of antioxidant activity, with the highest record of 156.87 g AK kg −1 in 'Alaja Krupnaja'. The lowest values of 61.70 g AK kg −1 and 63.59 g AK kg −1 were detected in 'Discolor 'and 'Titan', respectively.
The antioxidant activity values differ significantly in the published data as well, reflecting the influence of the applied method. For example, samples of sweet rowanberry from Poland showed antioxidant activity values of 10.75 µmol g −1 FW after using FRAP; however, only 5.94 µmol Trolox g −1 FW when ABTS radicals were employed [22]. A high AOA value detected by DPPH of 62.09 µg mL −1 was reported in the fruit of Sorbus torminalis (L.) Crantz from Turkey by Kıvrak et al. (2014) [24]. Another study emphasized the significant influence of the extraction solution on the final AOA value detected by DPPH; it reached 32.31 mg mL −1 after usage of the methanolic extract solution contrasting to 5.69 mg mL −1 after the application of the water extraction [41].

Influence of Various Factors on Antioxidant Activity
The method of regression analysis was applied to determine the correlation between the results for AOA, which were analyzed using three different methods: DPPH, ACW and ACL. The influence of TPC, TFC, AC and vitamins C and E on the AOA values was examined as well. The correlation coefficients are provided in Table 10. The relation between AOA determined by the DPPH method and TPC differs between studies. Kampuss et al. (2008) determined a considerable positive correlation only between the antioxidant activity and TPC (R = 0.886) [27]. In contrast with the results of this study, a negative correlation between AOA and TPC was reported in other studies [3,42].
This study showed a weak correlation between three methods of AOA detection: a positive direct correlation between DPPH and ACW (R = 0.3987), a very low correlation between DPPH and ACL (R = 0.0232) and an indirect correlation between ACW and ACL (R = −0.1965). A positive correlation of R = 0.977 was reported between DPPH and ABTS in the samples of hybrids of sweet rowanberry from Turkey, and indirect correlations between FRAP and DPPH (R = −0.703) and between FRAP and ABTS (R = −0.837) [41].

Influence of Total Phenolic Content (TPC)
Regarding the relation between TPC and ACW, a direct linear correlation with the highest value of R = 0.7671 was detected. Between TPC and DPPH, a very low correlation (R = 0.0596) was identified. Furthermore, the relation between TPC and ACL responded with an indirect, very low correlation of R = −0.2162. In the cultivars of sweet rowanberry from the Czech Republic harvested in 2011 and 2012, a high value of correlation between TPC and DPPH (R = 0.8904) was recorded [6]. These differences in correlations within phenolic groups and methods of AOA determination, despite using the same extracts, may be explained by different reaction mechanisms in the ORAC, ABTS •+ and DPPH • analyses [21].
The AOA values of the fruit extracts of S. torminalis were significantly influenced by the total phenolic content (TPC) established by the Folin-Ciocalteu method [9].
Notably variable correlations were reported between TPC and different methods for AOA detection in fruits of sweet rowanberry from Turkey: an indirect correlation between TPC and DPPH with a value of R = −0.728 and between TPC and ABTS with R = −0.855, contrasting with the positive linear correlation (R = 0.999) monitored between TPC and FRAP [41]. Similarly, in Sorbus samples from Poland, correlations between TPC and FRAP with a value of R = 0.993 and between TPC and ABTS with R = 0.943 were established [22]. In the six cultivars of Sorbus from Finland, a positive correlation (R = 0.868) was reported between TPC and FRAP, contrasting with the indirect correlation (R = −0.893) between TPC and DPPH [3].  [6]. Correspondingly, a considerable correlation between TFC and AOA in Sorbus fruits was observed by Kähkönen et al. (2001) [43].
A correlation between PCL and ACL (R = 0.4488) was established in contrast to the very low negative correlation between ACL and TFC.

Influence of Anthocyanin Content (AC)
The increased anthocyanin content in the cultivated species caused an insignificant rise in antioxidant activity [8].
The values of AOA detected by DPPH, ACW and ACL correlated with AC in various ways. A strong linear correlation between DPPH and AC was established, with R = 0.7132 representing the highest recorded value. Between AC and ACW and also AC and ACL, indirect negative correlations of R = −0.2227 and R = −0.0214 were determined, respectively. A high diversity between AC and FRAP/DPPH was published in the samples from Finland: between FRAP and AC with a value of R = 0.470 and between DPPH and AC with a value of R = −0.165 [3] Statistically significant correlations were reported between total flavonol content and antioxidant activity in sweet rowanberry cultivars from Finland [43]. Table 11 shows the values of the correlations between DPPH, ACW, ACL and the content of the flavonol rutin (RU), and the individual flavanols EGC, EC and C and total flavanols (FLAVAN) and stilbene resveratrol (RES). It is evident that flavanol catechin (C) mainly contributed to AOA as determined by the DPPH method (R = 0.8624). A direct linear correlation was estimated for epigallocatechin (EGC) with R = 0.2852 and for total flavanols (FLAVAN) with R = 0.6155. Epigallocatechin (EGC) and flavonol rutin (RU) displayed an indirect correlation with negative correlation coefficients of R = −0.4512 and R = −0.2054, respectively. Various correlations were identified with regard to the AOA detected by PCL. Between RU and ACW, a direct correlation of R = 0.6032 was determined contrasting with a weak indirect correlation between RU and ACL with R = −0.0595. Flavanols EC and ACW displayed an indirect correlation of R = −0.6050, in contrast to flavanols EGC, C and total flavanols (FLAVAN), which showed a weak direct correlation. The relation between ACL and individual flavanols showed direct linear correlations (R = 0.4562 for EGC, R = 0.2042 for EC and R = 0.3741 for K). The correlation between ACL and total flavanols (FLAVAN) was determined at R = 0.4885.
Between stilbene RES and ACW, a very strong linear correlation of R = 0.7169 was established, contrasting with the negative correlation of R = −0.6560 determined between RES and ACL. A very weak indirect correlation was detected between RES and DPPH (R = −0.1611).
Regarding phenolic acids, a considerable positive correlation was established, especially between hydroxycinnamic acid and antioxidant activity, in the samples of interspecific hybrids of sweet rowanberry from Finland [43].
Correlations between DPPH, ACW and ACL with the content of individual phenolic acids (GA, VA, SI, PK, PKEE, HB, EL, KA, FER, CHL, KU and SP) as well as with the total content of benzoic acid derivatives (DBA) and cinnamic acid derivatives (DCA) were determined by regression analyses. Table 12 provides their R values. As can be seen, DCA contributed to the AOA detected by DPPH to a greater extent (R = 0.7865) than DBA did (R = 0.4083). On the other hand, Sarv et al. (2021) identified weak correlations between different methods of AOA determination, including ORAC, ABTS •+ and DPPH, and HCA contents in the pomace extracts from sweet rowanberry [21].
DBA and ACW displayed a higher correlation value of R = 0.8478 than ACW and DCA (R = 0.3289). What is more, DBA and ACL showed a very weak indirect correlation (R = −0.0392) and DCA and ACL showed a very weak linear correlation (R = 0.0877). High variability was also reported in the correlations between DCA and FRAP (R = 0.070) and DPPH (R = −0.205) in the samples from Finland [3].
Regarding individual phenolic acids from the DBA group and DPPH, direct linear correlations were established (except for EL with R = −0.4118). The highest values of correlation coefficients were determined between DPPH and the acids PK and SI, with R = 0.6812 and R = 0.5412, respectively.
The highest value of R = 0.8014 displayed a correlation between DPPH and NCHA, followed by CHA (R = 0.6697). SP, with a value of R = 0.5326, showed a positive correlation with DPPH as well. PCA proved a very weak correlation (R = 0.0230). FEA displayed an indirect correlation with DPPH (R = −0.5284) and a very weak indirect correlation with CA (R = −0.0403).
Similarly, individual phenolic acids belonging to the DBA group and ACW showed direct correlations (except for EL with R = −0.4447). The highest values of the correlation coefficients of R = 0.8541, R = 0.6681 and R = 0.5314 were detected between ACW and HB, SI and PK acids, respectively. With regard to the acids belonging to the DCA group, the strongest linear correlations were determined between ACW and CA and FEA, with values of R = 0.7160 and R = 0.7032, respectively. The rest of the phenolic acids from the DCA group (CHA, NCHA and SA) showed only weak linear correlations, and PCA displayed an indirect negative correlation (R = −0.5179).
Regarding the relations between ACL and individual DBA acids, direct linear correlations were determined with these four acids: the strongest correlation was displayed between ACL and GA and VA (R = 0.6063 and R = 0.5961, respectively), EL and SI showed only weaker correlations with R = 0.4468 and R = 0.3461, respectively. An indirect linear correlation was established between ACL and PCEE, with the highest negative value of R = −0.6782, and a weaker correlation between ACL and HB and PC, with R = −0.2809 and R = −0.1143, respectively. Individual phenolic acids from the DCA group and ACL showed the strongest correlation (R = 0.5476). NCHA and CHA with ACL performed weak linear correlations, with values of R = 0.2026 and R = 0.1845, respectively. FEA, CA and PCA provided indirect linear correlations of R = −0.5167, R = −0.3978, and R = −0.0912, respectively.
Šavikin (2017) compared the chemical composition of fruits of S. aria and S. aucuparia from various altitudes. Even though no correlation was detected between TPC, total proanthocyanidins, antioxidant activity and the growing site [10], notable TPC values were confirmed in S. aucuparia and higher proanthocyanidin content in S. aria [10].

Conclusions
This paper has examined seven Sorbus cultivars and established their total polyphenolic, flavonoid and anthocyanin contents, as well as the contents of individual polyphenolic compounds belonging to the group of phenolic acids and flavonoids. Additionally, it has determined their antioxidant activity by employing various methods and assessed the relations and mutual influences of these compounds on the resulting antioxidant activity, also showing the impact of the method used through the evaluation of the correlation coefficients.
The highest total phenolic content of 8307.4 mg kg −1 was recorded in 'Granatina' with the main contribution being provided by phenolic acid content (7001.7 mg kg −1 ) and total flavonoid content (1304.6 mg kg −1 ). The lowest total phenolic content of 857.4 mg kg −1 was recorded in 'Discolor'. Flavanols represented the most abundant group of flavonoids in Sorbus fruits; catechin was the second most frequent flavanol, with the highest content of 633.67 mg g −1 in 'Granatina'. Flavonols were represented by rutin and quercetin. 'Businka' displayed a significant vitamin E content of 4.77 mg kg −1 , and 'Alaja Krupnaja' had the highest vitamin C level (7.89 g kg −1 ).
These results have proven the valuable potential health and nutritional benefits of Sorbus fresh fruits and their promising role in the food processing industry. What is more, this study facilitates the selection of the most suitable cultivars for producers and consumers based on a diverse array of aspects.

Data Availability Statement:
The data presented in this study are available within the article.