Antioxidant Activities of Vaccinium vitis-idaea L. Leaves within Cultivars and Their Phenolic Compounds

Lingonberry leaves are the subject of numerous studies because of antioxidant properties, positive influence on the health and potential use in the prevention and treatment of chronic diseases. In this work, the radical scavenging, reducing, chelating activities, and phenolic composition of ten lingonberry leaves cultivars, one subspecies, and one variety were investigated. Furthermore, the antioxidant activity of individual phenolic compounds, that can be found in lingonberry leaves, were analyzed, and structure-activity relationship was determined. Wide diversity for phenolic profile and antioxidant properties of lingonberry leaves has been observed in the present material. Cultivars ‘Kostromskaja rozovaja’, ‘Rubin’, and Vaccinium vitis-idaea var. leucocarpum surpassed all others tested cultivars and lower taxa by contents of phenolic compounds and antioxidant activity. Leaves of lingonberry cultivars and lower taxa are rich in arbutin, flavonol glycosides, proanthocyanidins, and the latter were considered to be the major contributor to antioxidant properties of lingonberry leaves.


Introduction
Lingonberry (Vaccinium vitis-idaea L.) is an evergreen shrub of the Ericaceae Juss. family. It is one of the most popular berries in Nordic countries, as well as in Baltic States, Russia and Canada [1,2]. Lingonberry raw materials are mainly collected in natural habitats. In Lithuania the area and productivity of the natural populations of lingonberries are declining, as well as they are unevenly distributed [3]. Research on natural resources assessment is relevant because anthropogenic factors have a negative impact on the populations of lingonberry. Sustainable use of natural resources becomes of crucial importance. Cultivation of introduced and locally selected lingonberries could be the best option to provide for the increasing needs of this plant material. Lingonberries are widely used in the human diet due to nutritional benefits, moreover, leaves of this plant can be used for prevention and treatment of urinary tract infections, stomach disorders, rheumatic diseases and hypercholesterolemia [1][2][3][4][5]. The wide spectrum of various biological properties of lingonberry leaves, including antioxidant activity is related to their phenolic constituents [4,6,7].

Ferric Reducing Antioxidant Power (FRAP)
The results provided in Figure 2 demonstrate the reducing activity of the different lingonberry cultivars and lower taxa. The reducing activity among the cultivars and lower taxa studied ranged with values from 743.38 to 1476.04 mg/g DW. The average of reducing power activity of different cultivars and lower taxa was 1101.32 mg/g DW. 'Rubin', 'Sussi' and 'Kostromskaja rozovaja' presented the highest reducing power activity and they did not differ significantly. These cultivars demonstrated significantly strong antioxidant activity compared to 'Erntedank' and 'Masovia' cultivars. Higher than average reducing activity was determined in V.vitis-idaea var. leucocarpum and 'Erntesegen' cultivar. Overall, 'Rubin' leaves revealed the greatest antioxidant properties, whereas the 'Erntedank' leaves possessed a very poor antioxidant activity-almost one and a half times lower than average.

Ferric Reducing Antioxidant Power (FRAP)
The results provided in Figure 2 demonstrate the reducing activity of the different lingonberry cultivars and lower taxa. The reducing activity among the cultivars and lower taxa studied ranged with values from 743.38 to 1476.04 mg/g DW. The average of reducing power activity of different cultivars and lower taxa was 1101.32 mg/g DW. 'Rubin', 'Sussi' and 'Kostromskaja rozovaja' presented the highest reducing power activity and they did not differ significantly. These cultivars demonstrated significantly strong antioxidant activity compared to 'Erntedank' and 'Masovia' cultivars. Higher than average reducing activity was determined in V. vitis-idaea var. leucocarpum and 'Erntesegen' cultivar. Overall, 'Rubin' leaves revealed the greatest antioxidant properties, whereas the 'Erntedank' leaves possessed a very poor antioxidant activity-almost one and a half times lower than average.

Ferric Reducing Antioxidant Power (FRAP)
The results provided in Figure 2 demonstrate the reducing activity of the different lingonberry cultivars and lower taxa. The reducing activity among the cultivars and lower taxa studied ranged with values from 743.38 to 1476.04 mg/g DW. The average of reducing power activity of different cultivars and lower taxa was 1101.32 mg/g DW. 'Rubin', 'Sussi' and 'Kostromskaja rozovaja' presented the highest reducing power activity and they did not differ significantly. These cultivars demonstrated significantly strong antioxidant activity compared to 'Erntedank' and 'Masovia' cultivars. Higher than average reducing activity was determined in V.vitis-idaea var. leucocarpum and 'Erntesegen' cultivar. Overall, 'Rubin' leaves revealed the greatest antioxidant properties, whereas the 'Erntedank' leaves possessed a very poor antioxidant activity-almost one and a half times lower than average.   Chelating activity showed statistically significant variations throughout the different cultivars and lower taxa of lingonberry leave extracts ( Figure 3). The ability of extracts to chelate ferrous ions ranged between 34.67 ± 3.05% and 79.85 ± 1.92%. The average of chelating activity of different cultivars and lower taxa of lingonberries was 62.21%. In this study, the highest chelating activity was determined in 'Rubin' cultivar, and it was almost the same as 10 µg/mL of standard chelator-EDTA (data not shown). There were no significant differences between the V. vitis-idaea var. leucocarpum and 'Rubin', 'Sussi', 'Kostromskaja rozovaja' cultivars. Higher than average chelating ability was also established in 'Kostromička', 'Erntesegen', and 'Entekrone' cultivars. This suggests that these cultivars of lingonberries are rich in compounds, which captures ferrous ions before ferrozine. Meanwhile, the lowest chelating activity-almost two times lower than average, was found in 'Koralle' cultivar, and one and a half times lower-in 'Erntedank' cultivar. By comparison with other cultivars, these differences were statistically significant. Obtained results indicate that there is a wide genetic diversity for chelating properties of lingonberry leaves, and chelating activity possibly depends on the phytochemical content of cultivar or lower taxa. Chelating activity showed statistically significant variations throughout the different cultivars and lower taxa of lingonberry leave extracts ( Figure 3). The ability of extracts to chelate ferrous ions ranged between 34.67 ± 3.05% and 79.85 ± 1.92%. The average of chelating activity of different cultivars and lower taxa of lingonberries was 62.21%. In this study, the highest chelating activity was determined in 'Rubin' cultivar, and it was almost the same as 10 µg/mL of standard chelator-EDTA (data not shown). There were no significant differences between the V.vitis-idaea var. leucocarpum and 'Rubin', 'Sussi', 'Kostromskaja rozovaja' cultivars. Higher than average chelating ability was also established in 'Kostromička', 'Erntesegen', and 'Entekrone' cultivars. This suggests that these cultivars of lingonberries are rich in compounds, which captures ferrous ions before ferrozine. Meanwhile, the lowest chelating activity-almost two times lower than average, was found in 'Koralle' cultivar, and one and a half times lower-in 'Erntedank' cultivar. By comparison with other cultivars, these differences were statistically significant. Obtained results indicate that there is a wide genetic diversity for chelating properties of lingonberry leaves, and chelating activity possibly depends on the phytochemical content of cultivar or lower taxa. The chelating activity of different cultivars and lower taxa of lingonberry leaves extracts; bars without the same letters (a, b, c, d, e, f, g, h) indicate statistically significant differences between the means (p < 0.05).

Qualitative and Quantitative Analysis of Phenolic Compounds in Lingonberry Leaves
The results of high-performance liquid chromatography-photodiode array detection (HPLC-PDA) method of analysis, showed that eleven phenolic compounds, belonging to subgroups of simple phenols, flavonols, flavanols, proanthocyanidins, and hydroxycinnamic acids were detected in different cultivars, and lower taxa of lingonberry leaves extracts (Table 1, Table 2). The chelating activity of different cultivars and lower taxa of lingonberry leaves extracts; bars without the same letters (a, b, c, d, e, f, g, h) indicate statistically significant differences between the means (p < 0.05).

Qualitative and Quantitative Analysis of Phenolic Compounds in Lingonberry Leaves
The results of high-performance liquid chromatography-photodiode array detection (HPLC-PDA) method of analysis, showed that eleven phenolic compounds, belonging to subgroups of simple phenols, flavonols, flavanols, proanthocyanidins, and hydroxycinnamic acids were detected in different cultivars, and lower taxa of lingonberry leaves extracts ( Table 1, Table 2).  Arbutin was the most abundant phenolic compound in all tested cultivars and lower taxa (Table 1). This compound contributed approximately 41-78% of the total phenolics in different cultivars and lower taxa of lingonberries. The coefficient of variation (CV) of arbutin among tested cultivars and lower taxa was 67%. Significantly the highest content of arbutin was found in 'Kostromskaja rozovaja', followed by 'Rubin' cultivar, meanwhile the lowest in 'Sanna', 'Erntedank', 'Erntesegen' cultivars of lingonberries.
The sum of detected hydroxycinnamic acids made up only 1-4% of the phenolic content in different cultivars and lower taxa of lingonberries ( Table 2). The CV was 55% and 59% for chlorogenic acid and cryptochlorogenic acid, respectively. Highest contents of chlorogenic and cryptochlorogenic acids were determined in the extracts from V. vitis-idaea var. leucocarpum.
Concerning the total amounts of detected phenolic compounds within the cultivars, it was determined that the greatest amounts of compounds were in the extracts from 'Rubin' and 'Kostromskaja rozovaja' cultivars, whereas the lowest ones-in 'Erntedank', 'Erntesegen', and 'Sanna' cultivars (data not shown).

Antioxidant Activity of Lingonberries Phenolic Compounds
In purpose of better understanding of structure-antioxidant activity relationship, we investigated not only standards of phenolic compounds, which we detect in our study but also several compounds which were detected in lingonberry materials in previous studies. The chemical structures of phenolic compounds, that can be found in lingonberry leaves, and have been investigated in our study, are shown in Figure 5.

Antioxidant Activity of Lingonberries Phenolic Compounds
In purpose of better understanding of structure-antioxidant activity relationship, we investigated not only standards of phenolic compounds, which we detect in our study but also several compounds which were detected in lingonberry materials in previous studies. The chemical structures of phenolic compounds, that can be found in lingonberry leaves, and have been investigated in our study, are shown in Figure 5. The radical scavenging activity of tested phenolic compounds and Trolox increased with their concentration. All calibration curves equations followed a linear regression model (p < 0.05) ( Table  3). Table 3. Calibration curves (y = ax + b) data obtained by ABTS assay for tested compounds.  The radical scavenging activity of tested phenolic compounds and Trolox increased with their concentration. All calibration curves equations followed a linear regression model (p < 0.05) ( Table 3). The Trolox equivalent antioxidant capacity values obtained by the ABTS assay (TEAC ABTS ) indicated how many times the radical scavenging activity of the tested phenolic compound was greater than that of Trolox ( Figure 6). Significantly the greatest TEAC ABTS values were determined for proanthocyanidins-procyanidins C1 and A2. Their scavenging activity was 2.92 ± 0.07 and 2.11 ± 0.03 times greater than Trolox, respectively. TEAC ABTS of monomeric flavanols was considerably lower than that of oligomeric polyphenols, mentioned above. The radical scavenging activity of (−)-epicatechin (1.71 ± 0.06) was greater than that of (+)-catechin (1.56 ± 0.02), and this difference was statistically significant. Among tested flavonols, quercetin was the most effective scavenger of the ABTS cation radical (1.89 ± 0.01). Glycosylation of the 3-hydroxyl in the quercetin molecule had an effect on the scavenging activity, as a result, TEAC ABTS of quercitrin, rutin, avicularin, and hyperoside was significantly inferior to quercetin. The scavenging activity of kaempferol glycoside-astragalin was lower than that of all quercetin glycosides, with the exception of quercitrin. Alkylation of catechol moiety further reduced ability to scavenge radicals, therefore TEAC ABTS of isorhamnetin-3-glucoside was lowest of all tested flavonoids (0.94 ± 0.05). Its TEAC ABTS was almost the same as that of glycosylated hydroquinone-arbutin (0.95 ± 0.02). The results showed, that hydroxycinnamic acids (ferulic, chlorogenic, and cryptochlorogenic acids) were less active than tested flavonoids, and despite structural differences, there was no statistically significant disparity among them.
was lower than that of all quercetin glycosides, with the exception of quercitrin. Alkylation of catechol moiety further reduced ability to scavenge radicals, therefore TEACABTS of isorhamnetin-3glucoside was lowest of all tested flavonoids (0.94 ± 0.05). Its TEACABTS was almost the same as that of glycosylated hydroquinone-arbutin (0.95 ± 0.02). The results showed, that hydroxycinnamic acids (ferulic, chlorogenic, and cryptochlorogenic acids) were less active than tested flavonoids, and despite structural differences, there was no statistically significant disparity among them. The radical scavenging activity of tested phenolic compounds; bars without the same letters (a, b, c, d, e, f, g, h, i, j) indicate statistically significant differences between the means (p < 0.05).

Ferric Reducing Antioxidant Power (FRAP)
It was determined that the reducing power of tested phenolic compounds and Trolox was concentration dependent. All calibration curves equations followed a linear regression model (p < 0.05) ( Table 4). The radical scavenging activity of tested phenolic compounds; bars without the same letters (a, b, c, d, e, f, g, h, i, j) indicate statistically significant differences between the means (p < 0.05).

Ferric Reducing Antioxidant Power (FRAP)
It was determined that the reducing power of tested phenolic compounds and Trolox was concentration dependent. All calibration curves equations followed a linear regression model (p < 0.05) ( Table 4). The Trolox equivalent antioxidant capacity values obtained by the FRAP assay (TEAC FRAP ) revealed how many times the reducing activity of the tested phenolic compound was greater than that of Trolox (Figure 7). The profile of reducing activity was similar to the results of the free radical scavenging activity. As in the ABTS assay, proanthocyanidins showed the greatest activity. TEAC FRAP was 2.97 ± 0.07 and 2.04 ± 0.04 for procyanidins C1 and A2, respectively. These results demonstrate that proanthocyanidins have a strong potency to donate electrons to reactive free radicals, converting them into more stable forms. In this assay there was no statistically significant difference in reducing activity between different isomers of catechin. As expected, quercetin was the most active flavonoid (1.76 ± 0.02). High TEAC FRAP was also observed in some quercetin glycosides-rutin (1.43 ± 0.05) and hyperoside (1.27 ± 0.05). The lowest reducing power had arbutin-its ability to reduce ferric ions was about three times lower than Trolox, and even eight and a half times lower than procyanidin C1. It was assessed that TEAC FRAP was 0.60 ± 0.01, 1.13 ± 0.02, and 0,86 ± 0.02 for ferulic acid, chlorogenic acid, and cryptochlorogenic acid, respectively.
Molecules 2019, 24, x FOR PEER REVIEW 12 of 22 Figure 7. The reducing activity of tested phenolic compounds; bars without the same letters (a, b, c, d, e, f, g, h, i, j) indicate statistically significant differences between the means (p < 0.05).

Ferrous Ions Chelating (FIC) Activity
In the ferrous ions chelating assay, we did not obtain the linear concentration-response curves, because results of simple linear regression analysis showed that there was no linear relationship between concentrations of phenolic compounds (tested concentration range 5-80 µg/mL) and chelating activity (p > 0.05 for regression coefficients). Therefore we only investigated and compared the chelating activity of the same 20 µg/mL concentration of the standards of phenolic compounds. All tested phenolic compounds were able to chelate ferrous ions at this concentration, nevertheless, they possessed significantly lower chelating activity than the same concentration chelating agent-EDTA (88.55 ± 2.84%). The greatest ability to chelate ferrous ions was observed in quercetin followed by procyanidin C1 (62.74 ± 1.80% and 58.53 ± 1.73%, respectively) ( Figure 8). Rutin was the most active quercetin glycoside, its chelating activity was similar to procyanidin A2 (48.38 ± 0.91% and 45.25 ± 1.86%, respectively). It was determined, that the type of other glycosides had no significant influence on the chelating activity. Furthermore, the chelating activity of all quercetin glycosides, except rutin, did not differ significantly from kaempferol glycoside and flavanols. The highest influence on chelating activity had methylation of catechol moiety, hence isorhamnetin-3-Oglucoside had the lowest activity of all tested compounds (12.08 ± 1.55%). Surprisingly, high chelating activity was determined in arbutin (33.44 ± 1.63%). We hypothesize that this is likely due to the specific pharmacophore and differences in reaction stoichiometry. By comparing the chelating activity of phenolic acids, a similar trend as in FRAP assay was found, with the exception, that there was no difference between the activity of chlorogenic acid isomers. Figure 7. The reducing activity of tested phenolic compounds; bars without the same letters (a, b, c, d, e, f, g, h, i, j) indicate statistically significant differences between the means (p < 0.05).

Ferrous Ions Chelating (FIC) Activity
In the ferrous ions chelating assay, we did not obtain the linear concentration-response curves, because results of simple linear regression analysis showed that there was no linear relationship between concentrations of phenolic compounds (tested concentration range 5-80 µg/mL) and chelating activity (p > 0.05 for regression coefficients). Therefore we only investigated and compared the chelating activity of the same 20 µg/mL concentration of the standards of phenolic compounds. All tested phenolic compounds were able to chelate ferrous ions at this concentration, nevertheless, they possessed significantly lower chelating activity than the same concentration chelating agent-EDTA (88.55 ± 2.84%). The greatest ability to chelate ferrous ions was observed in quercetin followed by procyanidin C1 (62.74 ± 1.80% and 58.53 ± 1.73%, respectively) ( Figure 8). Rutin was the most active quercetin glycoside, its chelating activity was similar to procyanidin A2 (48.38 ± 0.91% and 45.25 ± 1.86%, respectively). It was determined, that the type of other glycosides had no significant influence on the chelating activity. Furthermore, the chelating activity of all quercetin glycosides, except rutin, did not differ significantly from kaempferol glycoside and flavanols. The highest influence on chelating activity had methylation of catechol moiety, hence isorhamnetin-3-O-glucoside had the lowest activity of all tested compounds (12.08 ± 1.55%). Surprisingly, high chelating activity was determined in arbutin (33.44 ± 1.63%). We hypothesize that this is likely due to the specific pharmacophore and differences in reaction stoichiometry. By comparing the chelating activity of phenolic acids, a similar trend as in FRAP assay was found, with the exception, that there was no difference between the activity of chlorogenic acid isomers.
Molecules 2019, 24, x FOR PEER REVIEW 13 of 22 Figure 8. The chelating activity of tested phenolic compounds at a concentration of 20 µg/mL; bars without the same letters (a, b, c, d, e, f, g) indicate statistically significant differences between the means (p < 0.05).

Correlation Analysis
Antioxidant activities of lingonberry leaves extracts were correlated to the amounts of their phenolic constituents. There was a moderate positive correlation between reducing activity and the contents of cryptochlorogenic acid (R = 0.609, p < 0.05), (−)-epicatechin (R = 0.593, p < 0.05), arbutin (R = 0.622, p < 0.05), and furthermore between chelating activity and the contents of cryptochlorogenic acid (R = 0.665, p < 0.05) and arbutin (R = 0.621, p < 0.05). No significant correlations were determined between antioxidant activity and contents of other tested compounds.
Correlations between methods of analysis measuring different antioxidant properties were also investigated. There was an overall strong positive correlation between FIC and FRAP assays measuring antioxidant properties in cultivars of lingonberry leaves extracts (R = 0.844, p < 0.05). The assessed TEACABTS of studied phenolic compounds were similar to TEACFRAP (R = 0.904, p < 0.05). Moreover, there were a strong correlations between FIC values and TEACABTS (R = 0.817, p < 0.05) either TEACFRAP (R = 0.766, p < 0.05). This indicates that there is a relationship between the mechanism of action of phenolic compounds, and the same structural fragments may be responsible for different antioxidant properties.

Discussion
Phytochemical composition of lingonberry leaves strongly depends on the growing region. In our study, we have determined the quantitative and qualitative composition of bioactive compounds in lingonberries, the predominant compounds in phytochemical and antioxidant profiles, the differences in bioactive compound complexes of the distinct cultivated cultivars, subspecies, and variety.
Antioxidant properties of lingonberry leaves are in agreement with previous studies. Bujor et al. compared antioxidant activity between different parts of lingonberry, and determined that the antioxidant activity of lingonberry leaves is considerably greater than that of fruits [4]. Enkhtuya et al., investigated radical scavenging and reducing activities of different plants, and found that in Figure 8. The chelating activity of tested phenolic compounds at a concentration of 20 µg/mL; bars without the same letters (a, b, c, d, e, f, g) indicate statistically significant differences between the means (p < 0.05).

Correlation Analysis
Antioxidant activities of lingonberry leaves extracts were correlated to the amounts of their phenolic constituents. There was a moderate positive correlation between reducing activity and the contents of cryptochlorogenic acid (R = 0.609, p < 0.05), (−)-epicatechin (R = 0.593, p < 0.05), arbutin (R = 0.622, p < 0.05), and furthermore between chelating activity and the contents of cryptochlorogenic acid (R = 0.665, p < 0.05) and arbutin (R = 0.621, p < 0.05). No significant correlations were determined between antioxidant activity and contents of other tested compounds.
Correlations between methods of analysis measuring different antioxidant properties were also investigated. There was an overall strong positive correlation between FIC and FRAP assays measuring antioxidant properties in cultivars of lingonberry leaves extracts (R = 0.844, p < 0.05). The assessed TEAC ABTS of studied phenolic compounds were similar to TEAC FRAP (R = 0.904, p < 0.05). Moreover, there were a strong correlations between FIC values and TEAC ABTS (R = 0.817, p < 0.05) either TEAC FRAP (R = 0.766, p < 0.05). This indicates that there is a relationship between the mechanism of action of phenolic compounds, and the same structural fragments may be responsible for different antioxidant properties.

Discussion
Phytochemical composition of lingonberry leaves strongly depends on the growing region. In our study, we have determined the quantitative and qualitative composition of bioactive compounds in lingonberries, the predominant compounds in phytochemical and antioxidant profiles, the differences in bioactive compound complexes of the distinct cultivated cultivars, subspecies, and variety.
Antioxidant properties of lingonberry leaves are in agreement with previous studies. Bujor et al. compared antioxidant activity between different parts of lingonberry, and determined that the antioxidant activity of lingonberry leaves is considerably greater than that of fruits [4]. Enkhtuya et al., investigated radical scavenging and reducing activities of different plants, and found that in DPPH, ABTS, and FRAP methods of analysis lingonberry leaves were more active than leaves of black currant, hawthorn, strawberry, and tall currant. Radical scavenging and reducing properties of lingonberry leaves were similar to that of seabuckthorn [16]. Vyas et al. found significant differences of radical scavenging and reducing activities of lingonberry leaves within wild clones and cultivars, and distinguished V. vitis-idaea subsp. minus with high antioxidant properties [7]. Tian et al. assessed that within 10 min the lingonberry leaves extracts scavenged 86.3 ± 2.5% of DPPH radicals, and had an extremely high ORAC activity [13]. This indicates that lingonberry leaves exhibit great ability to donate hydrogen atom or transfer single-electron. Concerning chelating activity, some researchers reported that lingonberry extracts showed weaker chelating activity than extracts from plants belonging to other families [16,17]. However, it is difficult to compare the antioxidant activity results obtained in the present study with the literature data, because of the differences in methods and expression of antioxidant activity.
Based on our findings of the qualitative and quantitative analysis, arbutin was the predominant phenolic compound in all cultivars and lower taxa ranging from 7070.25 ± 288.64 µg/g up to 56,968.56 ± 2325.73 µg/g DW. Arbutin has been reported in previous studies to be a major phenolic constituent of lingonberry leaves [18,19]. The HPLC-PDA study of arbutin content in lingonberry leaves has given the result of 5168 ± 480 µg/g DW [20], which is similar to the result obtained in 'Kostromskaja rozovaja' cultivar.
Large amounts of flavonols in lingonberries have been confirmed by several researchers. Liu et al. determined that flavonol glycosides were the second major group of phenolics in lingonberry leaves [1]. Tian et al. emphasized that amount of flavonol glycosides is one of the main difference of phenolic profile between lingonberry leaves and fruits and assessed that contents of flavonol glycosides are considerably greater in lingonberry leaves [21]. Scientific studies demonstrate that not only hyperoside, avicularin, quercitrin, and astragalin can be detected in lingonberry leaves but also quercetin aglycon, rutin, and 3-O-glycosides of quercetin, isorhamnetin, and kaempferol [18,19,[21][22][23].
(+)-Catechin, (−)-epicatechin, proanthocyanidins, primarily as A-type or B-type procyanidin dimers and trimers, have been reported in a number of publications to be widespread in lingonberry leaves [1,13,[23][24][25][26]. However, the determination of proanthocyanidins by HPLC is problematic because proanthocyanidins often occur in complexes, that makes their separation and detection difficult [22]. This suggests that there are not only procyanidins C1 and A2, which were detected in present material but also more types of proanthocyanidins.
Low contents of hydroxycinnamic acids obtained in our research were partly consistent with previous studies. Bujor et al. reported that hydroxycinnamic acids represent the less abundant group of phenolic compounds in the lingonberry leaves. However, they found higher levels of hydroxycinnamic acids, because the relative content of hydroxycinnamic acids and their derivatives was in the range of 6-14% in leaves [4]. The presence of caffeoylquinic acid isomers is in agreement with their detection in lingonberries in various studies [18,23]. Nevertheless, Tian et al. disclosed that caffeoylquinic acids were not predominant compounds in lingonberry extract [13]. Other phenolic acids in various studies were detected in the extracts of lingonberries, i.e., coumaroyl quinic acid isomers (3.81% of the total combined peak area of all compounds), p-coumaric acid (0.64%), caffeic acid (0.61%), caffeoyl-shikimic acid (0.18%), feruloyl quinic acid (0.04%) [22], and also hexoses of ferulic acid [13], and free ferulic acid [27].
After the analysis of phytochemical and antioxidant profiles of separate lingonberry cultivars and lower taxa, there were no determined relationships with the country of origin of lingonberries nor the conditions of plants growing in trial areas. The cluster analysis revealed that according to the composition of phenolic compounds, the clusters were related to the countries of origin, especially with German and Russian origin cultivars. The predominant phenolic markers in German cultivars ('Erntekrone', 'Erntesegen', and 'Erntedank') were hyperoside and quercitrin, while in the Russian cultivars-procyanidins A2 and C1. In the study of phenolic compounds, it has been shown that in plant phylogeny and in the systematics, the composition of aglycons is a conservative feature that is very little dependent on the conditions of growth and denotes the specificity of genus and lower taxa [28]. Based on this research mentioned above, the composition of the phenolic compounds and their dominant markers are currently applied in both plant genotype studies [29], as well as antioxidant activity studies of these tribal representatives [28,30].
According to the results of radical scavenging, reducing and chelating activities of Lingonberries phenolic compounds, proanthocyanidins, which include in their structure epicatechin units, can be characterized by exceptionally strong antioxidant properties. These results agreed with those of Muselík et al. [31] and Spranger et al. [32], who reported that proanthocyanidins have the highest antioxidant activity among all natural phenolic compounds. This may be due to more hydroxyl groups, which act as strong hydrogen donors, generating by itself stable intramolecular hydrogen bonds with semi-quinoid free radicals and o-quinines, and thus disrupt radical chain reaction [33]. It was also found that B-type trimeric proanthocyanidins possess greater antioxidant activity than A-type dimeric proanthocyanidins.
In our study, all phenolic glycosides had lower antioxidant activity than their corresponding aglycons.
It confirms that O-glycosylation reduces the antioxidant activity of phenolic compounds [11,34,35]. It is believed that 3-O-glycosylation increases torsion angle and interferes the planarity of rings, leading to suppression of antioxidant activity [21]. Furthermore, it was noticed that in present study substitution of methoxyl group had a strong impact for reducing and chelating activities, whereas had a weaker impact for the radical scavenging activity. This suggests that the presence of electron donating methoxyl group adjacent to the hydroxyl group alters the redox potential and might enhance the hydrogen availability, donating capacity and thus radical scavenging activity [36,37]. Grzesik et al. discovered that the scavenging activity of the ABTS cation radical of ferulic acid was even greater than that of chlorogenic acid, meanwhile ability to reduce ferric ions was weaker [38].
The importance of catechol moiety (ortho-dihydroxyl structure) in the ring B was emphasized by previous studies [9,12,34]. In the present study, it was observed that catechol moiety increases the antioxidant activity because all quercetin glycosides still showed higher radical scavenging and reducing activities than kaempferol glycoside. Moreover, in all methods of analysis, scavenging, reducing and chelating activities of flavonol-quercetin were greater than that of flavanols-(+)-catechin and (−)-epicatechin. This can be explained by the presence of 3-hydroxy-4-keto conformation together with a 2,3-double bond in quercetin structure, which ensures extensive electron delocalization and higher stability to the aroxyl radical [12,39]. This structural fragment is associated particularly with a substantial ferrous ions chelation [10], consequently, quercetin can be used as a chelating standard [40]. The highest chelating activity of quercetin observed in our work supports this approach.
Determination of antioxidant activity of optical isomers showed that stereoisomery in flavanols concerning the position of the 3-hydroxyl group was not significantly important for reducing activity and ferrous ions chelation. This is in accordance with the findings reported earlier [10,31].
Greater reducing and chelating activity of chlorogenic and cryptochlorogenic acid compared with that of ferulic acid indicate that esterification of phenolic acids and the additional hydroxyl group on the aromatic ring have the significant influence on the antioxidant activity. These results are consistent with those of Natella et al. [37] and Sova [41], who reported that the antioxidant activity of phenolic acids is enhanced by the introduction of a second hydroxyl group. Chen et al. proved that esters can be an even more effective than free phenolic acids, suggesting that ester functional group favorably stabilize the specific radical species, and thus enhance the radical scavenging activity [42]. In addition, results from our study revealed that the 3-hydroxyl site on the quinic acid ring is more important for reducing activity than the 4-hydroxyl site.
The majority of the tested compounds in the ABTS and FRAP assays possessed higher free radical scavenging and reducing activities than Trolox (TEAC ABTS > 1 and TEAC FRAP > 1, respectively). Scientific studies confirm that natural phenolic compounds have similar or even greater antioxidant activity by comparison with Trolox [9,38,43]. However, all tested compounds in FIC assay had lower chelating activity than EDTA. This suggests that the chelating activity of phenolic compounds is slightly weaker than free radical scavenging and reducing activities.
The overall results of antioxidant activity of tested phenolic compounds indicate that procyanidins C1 and A2 are the major contributors to the antioxidant properties of lingonberry leaves extracts. This explains why extracts from 'Kostromskaja rozovaja', 'Rubin' cultivars, and V. vitis-idaea var. leucocarpum, containing the high procyanidins A2 and C1 content, possessed high reducing and chelating activities. Definitely, the antioxidant activity of the latter cultivar was predetermined by the high contents of flavonol glycosides and hydroxycinnamic acids.
Notwithstanding the fact that arbutin is more related for other biological properties, such as a diuretic, urinary antiseptic, melanin biosynthesis in human skin suppressing [6,18,20], and weak radical scavenging and reducing activities were obtained in the present study, correlation analysis still showed that this compound is related to antioxidant properties of lingonberries. So we assume that extraordinarily high contents of arbutin in all tested cultivars certainly contributes to total antioxidant activity.
One of the weakest antioxidant activity of 'Erntedank' cultivar can be explained by low contents of all detected phenolics, especially by small amounts of procyanidins A2 and C1. Low amounts of phenolics were also detected in extracts of 'Erntesegen' cultivar, however, interestingly great antioxidant activity, especially radical scavenging properties, were found for this cultivar. We hypothesize that this might be due to the presence of some other compounds and non-phenolic antioxidants, such as organic acids or triterpenoids, that can be found in lingonberry leaves [2,18,44]. Therefore, further research of lingonberry leaves composition is required.
peaty and well-drained soil. Fertilization and irrigation were applied during the season according to the meteorological situation. Plant raw material for research was collected in September 2017 and dried in 40 • C.

Preparation of the Lingonberry Leaves Extracts
The lingonberry leaves extracts were prepared using 0.25 g of dried raw material and 25 mL of 60% ethanol (1:100). The samples were extracted in an ultrasonic bath (Elmasonic P, Singen, Germany) at 25 • C for 15 min. The extracts were filtered through a 0.22 µm pore size filter (Carl Roth GmbH, Karlsruhe, Germany).

Preparation of the Phenolic Compounds
The stock solutions of phenolic compounds were prepared in 96% ethanol. For the ABTS and FRAP assays, the stock solutions were diluted to five concentrations (concentration ranges shown in Tables 3 and 4). For the FIC assay, all stock solutions were diluted to final 20 µg/mL concentration.

Determination of Antioxidant Activity
Antioxidant activity of the lingonberry leaves extracts and standards of phenolic compounds, which are most commonly found in lingonberry leaves, was analyzed by spectrophotometric ABTS, FRAP, and FIC assays.

ABTS Radical Cation Decolorization Assay
The radical scavenging activity was evaluated by the scavenging of ABTS radical cation as described by Re et al. with some modifications [46]. ABTS aqueous solution was mixed with K 2 S 2 O 8 to obtain a concentration of 2 mM. The stock solution was left for 16 h in the dark at room temperature. The working solution was then prepared by diluting the stock solution with water to obtain an absorbance of 1.0 at 734 nm against the blank (water). A volume of 20 µL of each extract (diluted 20 fold) or phenolic compound (at various concentration) was mixed with 3 mL of ABTS working solution. After 1 h, the decrease of absorbance was measured at 734 nm.
ABTS radical cation scavenging activity of extracts was expressed as antioxidant Trolox equivalents (TE) per gram of material. Calculated as follows: where: c-Trolox concentration in mg/mL from the calibration curve; V-the volume in mL; m-the exact weight of the dry material, g. The obtained absorption decrease of five different concentrations of tested phenolic compounds and Trolox were used to construct calibration curves (Table 3). Scavenging activity of the tested phenolic compounds was expressed according to Koleva et al. as TEAC ABTS [47], and calculated using the following equation: where: a sample -the slope of the sample obtained from the calibration curve by the ABTS assay; a trolox -the slope of Trolox obtained from the calibration curve by the ABTS assay.

Ferric Reducing Antioxidant Power (FRAP) Assay
The ability to reduce ferric ions was measured using a modified version of the method described by Benzie and Strain [48]. Briefly, 20 µL of each extract (diluted 20 folds) or phenolic compound (at various concentration) was added to 3 mL of freshly prepared working FRAP reagent consisted of a 10:1:1 ratio of solutions: 300 mM of acetate buffer (pH 3.6), 10 mM TPTZ dissolved in 40 mM HCl, and 20 mM FeCl 3 × 6 H 2 O. The absorbance of the mixture was measured at 593 after 1 h incubation at room temperature. The sample substituted by solvent was used as a blank.
Ability to reduce ferric ions by the extracts was expressed as antioxidant Trolox equivalents (TE) per gram of material. Calculated as follows: where: c-Trolox concentration in mg/mL from the calibration curve; V-the volume in mL; m-the exact weight of the dry material, g. The obtained absorption of five different concentrations of tested phenolic compounds and Trolox were used to construct calibration curves (Table 4). Reducing activity of the tested phenolic compounds was expressed as TEAC FRAP [47], and calculated using the following equation: where: a sample -the slope of the sample obtained from the calibration curve by the FRAP assay; a trolox -the slope of Trolox obtained from the calibration curve by the FRAP assay.

Ferrous Ions Chelating (FIC) Activity Assay
The chelation of ferrous ions was evaluated by the method according to Ye et al. [49], Zengin and Aktumsek [50] with slight modification. A volume of 1 mL of sample solution (extract diluted 3 fold or 20 µg/mL of the phenolic compound) was mixed with 50 µL of 2 mM FeCl 2 solution. After 5 min, the reaction was initiated by the addition of 0.2 mL of 5 mM ferrozine solution. The mixture was shaken and allowed to react at room temperature for 10 min. The absorbance of the solution was thereafter measured at 562 nm against the blank. To correct for background absorbance, a blank was prepared for each measurement by replacing ferrozine to solvent. The sample substituted by solvent was used as a negative control, and strong chelating agent-EDTA in a range 0-320 µg/mL was used as a positive control. The ability of the sample to chelate ferrous ions was calculated using the following equation: Ferrous ions chelating activity (%) = 1− As Ac × 100% (5) where: A s -absorbance in the presence of the sample; A c -absorbance of negative control.

Statistical Analysis
Statistical analysis was performed using SPSS 21.0 (SPSS Inc., Chicago, IL, USA) and Microsoft Office Excel 2010 (Microsoft, Redmond, WA, USA). All measurements were made in triplicate, and results were expressed as mean ± standard deviation (SD). Simple linear regression analysis was performed to calculate the concentration-response relationship of each investigated compound by ABTS, FRAP and FIC assays. Correlations were tested by using the Pearson correlation test. The hierarchical cluster analysis using Euclidean distances was performed. One-way analysis of variance was performed by ANOVA test. Significant differences between means were determined by Tukey HSD multiple comparison test. The p-values less than 0.05 were considered statistically significant.

Conclusions
Lingonberry leaves could be a promising source of bioactive compounds with notable antioxidant activity. The determined fingerprint profiles are important tools to prove the authenticity of lingonberry products. Therefore, the identification of secondary metabolites could be very important in chemotaxonomical and phytogeographical aspect. 'Kostromskaja rozovaja', 'Rubin' cultivars and V. vitis-idaea var. leucocarpum, possessed the greatest reducing, radical scavenging and chelating activities. Procyanidins A2 and C1 could be proposed as markers of radical scavenging, reducing and chelating activities of lingonberries.

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