Phenolic Profiles, Antioxidant Activity and Phenotypic Characterization of Lonicera caerulea L. Berries, Cultivated in Lithuania

Lonicera caerulea L. is an early fruit-bearing plant that originates from harsh environments. Raw materials contain a body of different phenolic origin compounds that determine the multidirectional antioxidant and pharmacological activities. The aim of this study was to comprehensively evaluate the phenolic composition, antioxidant capacities, vegetative, pomological, and sensory properties and their interrelations of selected L. caerulea cultivars, namely ‘Amphora’, ‘Wojtek’, ‘Iga’, ’Leningradskij Velikan’, ‘Nimfa’, ‘Indigo Gem’, ‘Tundra’, ‘Tola’, and fruit powders. Combined chromatographic systems were applied for the qualitative and quantitative profiling of 23 constituents belonging to the classes of anthocyanins, flavonols, flavones, proanthocyanidins, and phenolic acids. The determined markers of phytochemical profiles were cyanidin-3-glucoside, rutin, chlorogenic, and 3,5-dicaffeoylquinic acid. Anthocyanins and the predominant compound, cyanidin-3-glucoside, were the determinants of antioxidant activity. Cultivars ‘Amphora’, ‘Indigo Gem’, and ‘Tundra’ contained the greatest total amounts of identified phenolic compounds. Phenotypic characterization revealed the superiority of cultivars ‘Wojtek’ and ’Tundra’ compared to other cultivars, although ’Wojtek’ had low phenolic content and antioxidant activity and ’Tundra’ got lower sensory evaluation scores. Coupling the results of phenotypic and phytochemical characterization, cultivar ‘Tundra’ could be suitable for commercial plantations.


Plant Material
Vegetative growth parameters, yield, and fruit quality of eight Lonicera caerulea cultivars originated from Canada, Poland and Russia were investigated at the Institute of Horticulture, Lithuanian Research Centre for Agriculture and Forestry (Table 1) (55 • 4 55.67", 23 • 47 53.99" (World Geodetic System)). The plantation was established in 2016 on black woven mulch, planting distances were 3 × 1 m. A total of 10 plants of each cultivar were planted under a full randomized scheme. Vegetative growth was evaluated by measuring shrub height (cm) and width (cm). Shrub density was evaluated on the 5-point scale, where 1-very sparse; 5-very dense. Shrub health status was evaluated on the 5-point scale, where 1-dying; 5-excellent status. Annual yield (kg) per individual shrub was recorded and accumulated yield (kg) during 2018-2020 is presented as an average of ten shrubs. Berries were harvested when their color was uniform and berries easily separated from the stalk. Average berry weight (g) was measured of the sample of 100 berries. The three-year average weight is presented. Berry shape was established according to descriptors. Fruit sensory evaluation was done by trained panelists in 2020. Average scores of 7 evaluations are presented.

Preparation of L. caerulea Extracts
Fruits were collected and immediately subjected to freeze-drying in a Zirbus lyophilizer (Zirbus Technology GmbH, Bad Grund, Germany) at 0.01 mbar pressure and −85 • C condenser temperature. The dried fruits were milled to powder and kept in a sealed container in a dark dry place. About 1 g (precise weight) of freeze-dried L. caerulea powder was weighted in a dark glass vial, with 10 mL of 80% (v/v) ethanol acidified with 2% hydrochloric acid. The extraction process continued for 40 min at 80 Hz and 904 W, in an ultrasonic bath (Elmasonic P, Singen, Germany). The extracts were filtered through 0.22 µm pore size membrane filters (Carl Roth GmbH, Karlsruhe, Germany) and transferred to the dark glass vials.

Antioxidant Activity Assays
The ABTS assay was performed as described by Re et al., 1999 with some modifications [24]. Briefly, 3 mL of diluted ABTS radical cation solutions, produced by reacting 7 mM ABTS aqueous solution with 2.45 mM potassium persulfate and allowing the mixture to stand for 16 h in dark, were mixed with 20 µL of extracts. The decrease in absorbance was recorded at 734 nm after 1 h of incubation.
The ferric reducing activity (FRAP) was determined according to the method of Benzie and Strain (1996) [25]. During the evaluation, 3 mL of freshly prepared solutions of FRAP reagent, consisting of 300 mM acetate buffer, 10 mM TPTZ in 40 mM HCl, and 20 mM iron (III) chloride in a final ratio of 10:1:1 (v/v/v) were mixed with 20 µL of extracts, following by incubation for 1 h and absorbance recording at 593 nm.
The CuPRAC (cupric-reducing antioxidant capacity) assay was performed as described by Apak et al., 2007 [26]. In this assay, 3 mL of solutions of CuPRAC, consisting of 0.01 M copper (II) chloride, 0.001 M ammonium acetate buffer solution, and 0.0075 M neocuproine in a final ratio of 1:1:1 (v/v/v), were mixed with 20 mL of extracts, and the absorbance was measured at 450 nm.
All antioxidant activity measurements and calculations were performed using Trolox calibration curves and were expressed as µmol of the Trolox equivalent (TE) per one gram of dry weight, according to our previous research [27].

Statistical Analysis
Statistical analysis was performed using IBM SPSS 24.0 (SPSS Inc., Chicago, IL, USA) and Microsoft Office Excel 2017 (Microsoft, Redmond, WA, USA). All measurements were performed in triplicate, and results were expressed as mean ± standard deviation (SD). The one-way analysis of variance was performed, and post-hoc Tukey HSD multiple comparison test was used to identify significant differences at p ≤ 0.05. Radical scavenging and reducing activities were expressed as Trolox equivalent antioxidant activity as mean ± SD. Principal component analysis (PCA) was performed considering factors with eigenvalues higher than 1. Regression analysis was performed for the calibration curves of concentration-response. Correlations were evaluated using the Pearson correlation coefficient.

Evaluation of the L. caerulea Cultivars
The blue honeysuckle shrubs still did not reach their final size in the 5th year after planting ( Table 2). Shrub height and width varied among cultivars from 87 up to 120 cm, and from 90 up to 110 cm, respectively. Though there were significant differences between cultivars both for shrub height and width, some more years are needed to draw final conclusions. The two cultivars 'Amphora' and 'Leningradskij Velikan' were distinguished by very dense shrubs; this is a negative character that bears difficulties during the berry harvest. Other tested cultivars did not differ significantly in shrub density. The two Polish cultivars, 'Iga' and 'Tola', were the same as the Canadian cultivar 'Tundra'; in their 5th growing season, they had very healthy shrubs without any visual decline symptoms. 'Leningradskij Velikan' had the worst health status and the main negative symptoms were stunted new growth and leaf discoloration. The highest cumulative yield was obtained of cultivar. 'Tundra' (Table 3). Cultivar 'Wojtek' also was very productive and did not differ significantly from cultivar 'Tundra'. A total of five kg of berries harvested during three years, from 3-5 year old shrubs, were higher than the yield reported from the trials in Poland, where 3-4 years old cultivar 'Wojtek' gave around 1 kg yield [31], though other authors claim that 2 kg of berries from a shrub is an average yield for 5-6 year old shrubs [32,33]. In our trial, berries were harvested in one pick though there are recommendations to harvest in 3-4 times, or even up to 7 times, to prevent loss of overripe berries [34]. To consider that conditions cumulative yields in our trial could be increased by 10-15%. Cultivar 'Amphora' was the lowest yielding cultivar (1.8 kg during three years), and only cultivar 'Nimfa' did not differ significantly from it. Average berry weight varied from 0.95 up to 1.19 g between cultivars (Table 2). Cultivars 'Iga' and 'Wojtek' had significantly larger fruits than most other cultivars tested. Such fruit size obtained in our trial is comparable to the results of some Polish and Slovenian trials [16,35].
Szot and Lipa (2013) reported a significant increase in average berry weight after the shrub pruning [36], but usually pruning starts from the 6th-8th year after planting of blue honeysuckle plantations, and our plants did not reach that age yet.
Sensory evaluation revealed significant differences between cultivars in berry appearance, flavor, and taste character ( Table 4). The most attractive were berries of 'Amphora', 'Wojtek' and 'Tola'. Berries of 'Iga' and 'Tundra' were evaluated to have a significantly lower total score. On the other hand, 'Leningradskij Velikan' berries had the best flavor score, possibly related to higher dominance of sweetness. The flavor score of the most attractive 'Amphora' and 'Tola' berries was the lowest, the same as 'Iga', which lead to lower overall evaluation. It is interesting that the taste character of all these cultivars was evaluated as sour or acid. Combining berry appearance and flavor, 'Leningradskij Velikan' had the highest ratings, whereas 'Indigo Gem', 'Nimfa' and 'Wojtek' did not differ significantly.

Phenolic Profiles of Fruits of Selected L. caerulea Cultivars
Phytochemical profiles are determined by genetic origin, harvesting, and processing techniques of the plant materials and the environmental growing conditions. Environmental conditions are one of the main detrimental factors affecting the qualitative and quantitative compositions of plant materials. Central Europe is ascribed to the region of favorable growing conditions for L. caerulea species, as well as Canada and Northern countries of origin [2,9]. The HPLC-PDA and UPLC-PDA assays enabled profiling and quantification of anthocyanins, flavones, flavonols, and hydroxycinnamic acid contents in selected cultivars (Figure 1). Anthocyanins were the prevailing compounds and constituted from 38 up to 91% of total identified compounds, flavonoid fraction comprised from 3 up to 36%, and phenolic acids-3-42%.
Rutin was the phytochemical marker of the flavonoids ranging from 255.78 ± 8.38 µg/g in 'Tola' up to 779.31 ± 9.88 µg/g in 'Indigo Gem' cultivar. The other quantified flavonoid derivatives were isoquercitrin, quercitrin, quercetin, isorhamnetin, luteolin-7-glucoside, and apigenin, showing significant quantitative variation between tested cultivars. The flavonoid profiles of L. caerulea consist of (+)-catechin, (-)-epicatechin, quercetin, isorhamnetin, kaempferol, apigenin, and glycosides and aglycones of luteolin [6,7,11]. Rutin is the predominant compound in the flavonoid fraction and can be regarded as the phytochemical marker. These results are in agreement with various studies regarding the chemical variation of flavonoids [7,11,13,37]. In our study, Russian cultivars contained the greatest amounts of flavonoids in the following order 'Nympha' > 'Amphora' > 'Leningradskij Velikan'. Gawronski et al., 2020 determined 'Aurora' with the highest amounts of flavonoids from the 30 tested L. caerulea cultivars [17]. Great variation in total contents of flavonoids was determined by various authors for the 'Indigo Gem', 'Leningradskij Velikan', 'Nimfa' cultivars due to different environmental conditions and growing techniques [17]. Anthocyanins and flavonoids are mainly located in skin tissues [6], therefore whole fruits or pomace materials could be selected for obtaining fruit powders for further functionalization [10].
Significantly, the greatest total amount of proanthocyanidins, determined by spectrophotometric DMCA method, were determined in samples of cultivars 'Amphora' and 'Leningradskij Velikan' (2.16 ± 0.03 mg/g and 2.13 ± 0.01 mg/g, respectively) ( Table 7). The amounts of procyanidin B1, determined by HPLC-PDA method, were well correlated with the total amounts of proanthocyanidins (R = 0.886; p < 0.05) and were highlighted in 'Amphora' and 'Leningradskij Velikan', while 'Tola' and 'Wojtek' contained the lowest amounts (Table 4). A much higher total of proanthocyanidins levels, compared to detected procyanidin B1 amounts, indicate that there are much more unidentified proanthocyanidins in tested L. caerulea samples. Procyanidin dimers, trimers and up to polymers has been detected in the fruit samples by various authors [1]. Significant negative correlational interdependence was also determined between the total proanthocyanidins and average fruit weight (R = −0.602; p < 0.05). Proanthocyanidins in L. caerulea are also geographic origin-specific [6]. Their quantities are determined by the genotype and are negatively associated with the ripening [4]. The procyanidin dimers, trimers, and tetramers were determined in different cultivars [7,10,11]. Kucharska et al. study confirmed a great variation of amounts of individual and total proanthocyanidins in different cultivars [7]. The results are comparable with our tested cultivars.
The identity of flavonoids and phenolic acids in extracts of L. caerulea fruits was additionally confirmed by mass spectrometry, which data are presented in Table 6. Obtained mass fragmentation spectra, m/z proportions were identical with MS/MS data of reference compounds and literature.

Antioxidant Activity of L. of Fruits of Selected L. caerulea Cultivars
Antioxidant activity of phenolic rich plant materials is highly correlated with the antioxidant capacity, which depends on the structural peculiarities of compounds [27,43]. In this study, the antioxidant activity was evaluated using the in vitro techniques that differ in mechanism of action and experimental conditions. The radical scavenging activity was evaluated using ABTS in neutral medium, while reducing activity-FRAP (pH is acidic) and CuPRAC (pH-close to physiological values) assays. ABTS, FRAP, and CuPRAC belong to the single electron transfer based assays [44]. CuPRAC assay due to the electronic configuration of copper complex possess faster kinetics compared to FRAP assay [26].

Principal Component Analysis of Fruits of Selected L. caerulea Cultivars
The principal component (PCA) analysis was performed on the phytochemical profile components, anthocyanins, flavones, flavan-3-ols, flavonols, proanthocyanidins, and hydroxycinnamic acids. The three main derived principal components explained 78.82% of the total variance. The score plot model has shown good separation between the investigated cultivars of L. caerulea. The PC1 was positively correlated with the amounts of identified anthocyanins, namely cyanidin (0.927), peonidin (0.956), cyanidin-3-glucoside (0.945), peonidin-3-glucoside (0.937), cyanidin-3-galactoside (0.915), cyanidin-3-rutinoside (0.836), cyanidin-3,5-diglucoside (0.799), and constituted 43.18%. The PC2 described 20.34% of the total variance and correlated positively with all the determined caffeoylquinic acids (0.675-0.918), procyanidin B1 (0.530). The PC3 accounted for 15.30% of the total variance and was well correlated with the amounts of procyanidin B1 (0.592) and flavonoid aglycones, namely apigenin, quercetin, (0.823, −0.958, respectively). The arrangements of score plots of investigated cultivars are shown in Figure 2. The first segregated group included 'Wojtek', 'Iga', and 'Tola' cultivars. They all were characterized by the lowest amounts of anthocyanins, proanthocyanidins, rutin, isoquercitrin, dicaffeoylquinic acids, and lowest or average reducing and radical scavenging activities. On the other hand, this group was distinguished with the greatest amounts of luteolin-7-O-glucoside, as well as, greatest fruit weight, and the highest score of appearance. The second group included 'Indigo Gem' and 'Tundra' cultivars. They possessed the greatest reducing activities in FRAP and above the average in CuPRAC assays, as well as high amounts of cyanidin-3-glucoside. Although, they had the lowest shrub height and fruit weight. Cultivars 'Leningradskij Velikan', 'Nimfa', and 'Amphora' tended to be specific. Cultivar 'Amphora' distinguished with the highest amounts of anthocyanins and flavonol derivatives. Samples of the 'Leningradskij Velikan' cultivar were the richest in procyanidins, and flavonoid aglycones, namely apigenin, isorhamnetin. Both cultivars possessed high shrub width and the greatest density. Cultivar 'Nimfa' can be characterized by the greatest amounts of individual caffeoylquinic acids. The origin place of the cultivar determines the phytogeographical profile in a qualitative and quantitative manner. Numerous studies support information, that Russian cultivars contain higher amounts of phytochemicals [3,4,17]. In our study, the PCA analysis on phenolics clearly defined the cultivars into groups in the relation to their origin. The lowest (p < 0.05) total amounts of all groups of identified compounds, as well as, the lowest antioxidant capacities were determined for the Polish cultivars, namely 'Wojtek', 'Iga' and 'Tola'. Genotype characterization resulted in the highest appearance and fruit weight scores. The Canadian cultivars 'Indigo Gem' and 'Tundra' were distinguished with the highest ferric reducing antioxidant power and low fruit weight. The Russian cultivars had the greatest anthocyanin contents and were specific in dominant phenolics of different chemical groups.