Relationship of colour with the phytocompounds present in Cornus mas cultivars

ABSTRACT Cornelian cherry (Cornus mas L.) fruits are a well-known source of antioxidants and other biologically active compounds, and these compounds depend on maturity. Ripeness is recognized by means of a visual assay of the growing fruit. The study aimed to determine the relationship between the color of the tested cultivars of Cornus mas and their content of the predominant compounds (i.e., flavonoids, anthocyanins, vitamin C, carotenoids and chlorophyll). The studied deviation between cultivars is based on the tested parameters and compared with the genetic profile of Cornelian cherry cultivars. Cornelian is a rich source of anthocyanins and flavonoids. Particularly valuable is the cv. Szafer variety, which color is the darkest and the richest in phenols, flavonoids and anthocyanins. However, the correlation between colorimetric and chemical parameters is found to be low for most variables. Genetic polymorphism analysis showed different distances between the cultivars than the values resulting from chemical analyses. The part of fruit, which was subjected to colorimetric measurement, did not affect the distance projection. Colorimetric parameters were poorly correlated with spectroscopic results, but did not change the distances between the samples. Differences in fruit color and composition, and genetic relationship can be drawn from the adaptation of different cultivars to the current study.


INTRODUCTION
Cornelian cherry is a plant growing in Eastern Europe, where fruits are used in traditional cuisine as a component of fruit preserves and liquors. Cornelian cherry fruits are rich in vitamin C and phenolic compounds, i.e., flavonoids, phenolic acids, and anthocyanins. [1] The latter are mainly responsible for the anti-proliferative and antioxidant properties of Cornelian cherry juice [2,3] . The maturity stage of Cornelian cherry fruits affects the levels of phytocompounds they contain, especially plant secondary metabolites (e.g., anthocyanins) (Szczepaniak et al. 2019a; Yarılgaç et al. 2019). Because Cornelian cherry is seasonal, there is a need to control the stage of ripeness. Chemical analyses are timeconsuming and expensive, therefore the aim of our study was to determine the relationship between the color of the ripe fruit and its chemical constituents.

Studied material
The ripe fruits of Cornelian cherry (Cornus mas L.) were collected from the orchard farm Szynsad in Dąbrówka Nowa, Błędów, Mazowieckie, Poland (51°47′01″N 20°43′04″E) in 2018. The fruit was harvested from the third year of plant cultivation on soils characterized by pH 6.1 and the content of humus 1.1%. Soil samples for chemical analysis were collected with a specialized soil auger. Nutrient contents (pH, humus) in the soil were determined at the District Chemical and Agricultural Station in Poznań.
The average amount of precipitation in the growing season equaled 320 mm/m 2 , with an average temperature of approx. 14.9°C. Characteristics of climatic conditions which prevailed during the field research were based on data from the meteorological station belonging to the farm (51°47′01″N 20°43′ 04″E), which the fruits for the study originated from. The fruits of five cultivars grown in Poland were used: Szafer, Jolico, Florianka, Słowianin, and P5. The fruits were stored in cooling conditions (4°C) until the extracts were prepared.

Extraction
Cornelian cherry extracts comprised fresh fruits according to the methodology described by Szczepaniak et al. [4] . The fruit skin was cut to allow proper maceration of the plant material. The extract used -40% (v/v) ethanol and the entire maturation process were chosen to simulate the tincture production process. The fruit matured into extractants at a ratio of 1 g fruit: 5 ml extractant. The average mass of the fruit used to obtain one extract approximated 26 g. Samples were encoded as follows: S-cv. Szafer extract P-cv. P5 extract; V-cv. Słowianin extract; F-cv. Florianka extract; and C-cv. Jolico extract.

Color measurements
The color parameters of Cornelian cherry's flesh (M), skin and these two anatomical parts together (M + O) were investigated. Fruit skins were measured both in the frozen (Z) and thawed (O) state. Color measurement was made in the L* a* b* CEN unit system using spectrometer CM-5 (Konica Minolta, Japan), accordingly to the methodology described by the device manufacturer. As a source of light, D 65 was applied, and the color temperature was 6504 K. The observation angle of the standard colorimetric observer was 10°. Measurements for each anatomical part and for each cultivar were repeated five times. The instrument was calibrated using a black pattern.

Total phenolic content
Total phenolic content (TPC) was determined using own spectroscopy methodology [4] . The total amount of 5 mml of the tested extract and the volume of Folin-Ciocalteau reagent (Chempur, Piekary Śląskie, Poland) were added to a 50 ml volumetric flask. After mixing the mixture for 5 s, 5 ml of supersaturated solution of Na 2 CO 3 and NaHCO 3 were added. Before the absorbance measurement, the sample was incubated in a dark room for 90 min. The analysis was carried out with a Specord 40 spectrophotometer (Analytic Jena, Germany), and the absorbance was measured at λ = 765 nm. Each sample was measured in a six-fold repetition and for each tested extract the whole analytical procedure was repeated three times. The calibration curve was based on gallic acid (Sigma-Aldrich, Germany) at five different concentrations: 22.21 mg/ml; 26.65 mg/ml; 31.09 mg/ml; 35.54 mg/ml and 44.42 mg/ml. Each standard solution was measured six times. The r 2 value of the calibration curve was 0.9523, and the curve was validated using internal cross validation. The total average recovery observed was 102.9%. The results are related to the fresh weight of the extract and expressed in milligrams of gallic acid equivalent (GAE)/100 g fresh weight (f.w.).

Total anthocyanin content
The content of anthocyanins was determined using the spectroscopic method described by Lee et al. [8] with slight modifications. The assay was based on measuring the difference in absorbance between the two wavelengths and buffer media in which the 0.4 ml test sample was diluted 10 times, i.e., either 25 mM KCl solution in 0.1 M HCl (pH = 1.0) or sodium acetate buffer (pH 4.5). Total anthocyanin content (TAC) was calculated using the following equation. The results were given in mg cyaniding-3ʹglucoside equivalent (CGE) per 100 g of fruit.
TAC ¼ Where: A 510 is the absorbance at λ = 510 nm; A 700 is the absorbance at λ = 700 nm. The measurement was repeated three times for each sample using the Specord S40 (Analytic Jena, Germany) spectrophotometer. Buffer solutions were used as blank samples.

Total flavonoid content
The total flavonoid content (TFC) was determined by a modified method of Meda et al. [9] based on the colorimetric reaction between flavonoids in the tested extract and aluminum ions. A 2% AlCl 3 solution was prepared in methanol and mixed in a test tube with the tested extract at a ratio of 1:1 (v/v). Then, after 10 minute conditioning in the dark place, absorbance was read at 415 nm. All extracts were analyzed in triplicate. A mixture of pure methanol (Honeywell, UK) and the tested extract (1:1 v/v) was used as a negative sample. The absorbance of the negative sample was subtracted from the Cornelian cherry extract. The standard curve was prepared using standard quercetin solutions (Sigma-Aldrich, Germany) (r 2 0.9989). The final results are given in mg quercetin equivalent (QE)/100 g f.w.

Total carotenoid and chlorophyll content
The assay was conducted according to the Abou-Arab et al. method. [10] The obtained extracts were subjected to multi-wavelength absorbance measurement in the SPECORD S40 apparatus (Analytic Jena, Germany). Absorbance was read for the following wavelengths: 440 nm (A 440 ); 644 nm (A 644 ) and 662 nm (A 662 ). Each extract was analyzed in triplicate. The content of chlorophyll A (CA), chlorophyll B (CB), and total carotenoids (TC) was given in mg/100 g f.w. and calculated according to equations (2)(3)(4). The measurements were made in fourfold repetitions.

Vitamin C determination
The content of vitamin C (AA) was determined according to Omaye et al. spectroscopy method, [11] modified independently in the following way. The whole volume of 600 μl of the tested extract was added to the test tube and then 300 μl of 0,85 M trisodium citrate (POCH, Poland) solution was added. Next, 590 μM of dichlorophenolindophenol (DCPIP, Sigma-Aldrich, Germany) was added, and after 30 s, absorbance (A 1 ) was measured at a wavelength of 520 nm (Metertek, Taiwan). After that, the absorbance of the tested solution was scavenged by adding ascorbic acid (AA, Chempur, Poland) to the cuvette. The sample with AA overdose was measured (A 2 ). The measurement was made in threefold repetition. In the negative control sample, 600 μl of 40% ethanol was used instead of the tested extract. The blank sample was the chosen extractant -40% ethanol. The absorbance drop (X) was calculated according to the following equation.
Where A n1 is the absorbance of the negative control; A n2 is the absorbance of the negative control after scavenging by ascorbic acid additive. The calibration curve was made using AA standard solutions. The linearity of the curve coefficient (r 2 ) equaled 0.98. The final results are given in mg AA/100 g f.w.

Phenolic acids and flavonoids quantitative determination
Phenolic compounds in samples were analyzed after performing alkaline and acidic hydrolysis. [12] The analysis was carried out using the Acquity H class UPLC system equipped with the Waters Acquity PDA detector (Waters, USA). Chromatographic separation was performed on the Acquity UPLC® BEH C18 column (100 mm×2.1 mm, particle size 1.7 μm) (Waters, Ireland). The elution was carried out in a gradient mode using the following mobile phase composition: A: acetonitrile with 0,1% formic acid, B: 1% aqueous formic acid mixture (pH = 2). Concentrations of phenolic compounds were determined using an internal standard at wavelengths λ = 320 nm and 280 nm and finally given at mg/100 g f.w. of Cornelian cherry. Compounds were identified based on a comparison of retention time of the analyzed peak with the retention time of the standard and by adding a specific amount of the standard to the analyzed samples and repeating the analysis. .

Genetic relationship analysis
The tested fruits were frozen in liquid nitrogen and pounded into powder in a porcelain mortar. DNA was isolated using the Junghans and Metzlaff's methods. [13] The quantity and quality of the isolated nucleic material were validated by the Nanodrop (ThermoFscher, UK) spectrometer at wavelengths 260, 280 and 230 nm. For the analyses, 14 pairs of SSR (Simple Sequence Repeats) markers and 34 ISSR (Inter Simple Sequence Repeat) markers (Table 1)  The polymerase chain reaction (PCR) was conducted in the Biometra thermocycler. PCR for SSR was conducted in 35 cycles and involved the stages: initial denaturation at 95°C for 5 min, denaturation at 94°C for 1 min, attaching of starters at 52°C for 30 s, amplification at 72°C for 30 s, final amplification at 72°C for 10 min. PCR for ISSR was conducted in 40 cycles and involved: initial denaturation at 95°C for 5 min, denaturation at 94°C for 1 min, binding of starters at 52°C for 30 s, amplification at 72°C for 30 s, final amplification at 72°C for 10 min. Amplification results were checked using capillary electrophoresis in the Qiaxcel apparatus (Qiagen, Hilden, Germany). The binary matrix was composed in the Qiaxcel ScreenGel software using the Binary Peak (Qiagen) function.

Statistical analysis
Statistical analysis was conducted using the Statistica 13 software (StatSoft, Poland). We treat color parameters (L*, a* and b*) as the former quantitative dataset. The latter comprised the other parameters discussed in this study. A cultivar was treated as a qualitative variable. We performed the principal component analysis of the studied parameters for the tested cultivars. Moreover, we showed correlations between the tested cultivars using the Euclid's tree projection. To construct an Euclid's tree projection for all tested parameters, the results of color measurements for different anatomical parts and cultivars were treated as separate variables.
The results of genetic polymorphism were analyzed statistically in the R software. [14] To illustrate the genetic distance between the tested cultivars, a distance tree chart was used based on the Unweighted Pair Group Method with Arithmetic Mean (UPGMAM).

Color measurements
The colorimetric assay showed significant differences between the analyzed Cornelian cherry fruits and their anatomical parts (Table 2). Flesh, considered individually and together with fruit skin, tended to have lower brightness L* values than fruit skin alone. Significant differences between the color of frozen and thawed fruit can be noticed in a* and b* values.
In the context of cultivars, similar values were observed for cultivars Słowianin and P5.

Total anthocyanin content
Cultivar Szafer was found to be the richest in anthocyanins (Table 3), similarly to the TFC test. The content of anthocyanins was approximately 10 times higher than in the P5 cultivar, which had the lowest content of anthocyanins. In contrast to the results observed for flavonoids, the samples for this test varied much stronger.

Ascorbic acid content
In contrary to the tests discussed above, the richest in AA was cv. Jolico, and the poorest was cv. Słowianin (Table 3). As in the case of total anthocyanin content, it found out that cultivars Slowianin and Florianka had similar values. The remaining cultivars differed significantly both from them and from each other.

Chlorophyll fractions and total carotenoid content
The tested Cornelian cherry cultivars differed significantly in the content of total chlorophyll, chlorophyll A, chlorophyll B, and total carotenoids ( Table 3). The richest in chlorophylls was cv. Szafer -7.21 ± 0.06 mg/100 g f.w., while cv. Florianka had the lowest total chlorophyll content -0.16 ± 0.01 mg/ 100 g f.w. Among the chlorophyll fractions, chlorophyll B dominated in all tested samples, and its content was approximately twice as high as that of chlorophyll A. Total carotenoid content ranged from 0.16 to 1.42 mg/100 g f.w. between samples, and no carotenoids were found in cv. P5 (Table 3).

HPLC determination of flavonoids and phenolic acids
The compounds detected in the tested Cornus mas extracts, can be considered as two separate groups ( Table 4). The first of the dominant compounds comprises chlorogenic acid, gallic acid, quercetin, rutin and naringenin, while the other one involves compounds, whose concentration is low or scarce (e.g., vanillic and salicylic acids). The PCA analysis (Figure 1 a) showed that most phenolic acids and flavonoids are correlated, as they are located in the same upper left corner of the plot. Exceptions are salicylic acid situated near AA and naringin adjacent to the point of cultivar Słowianin.

Correlation analysis
After including all the above parameters toward a single statistical model, two significant implications appeared. First, the anatomical part variable had no higher statistical significance for several samples than the issue of cultivar (Figure 2a). The Euclid's tree projection presents a similar arrangement of cultivars to that presented in Figure 1. The PCA analysis has confirmed such significance, as it is presented in the previous points. A close relationship between TFC, cv. Szafer and chlorophyll fractions was found. Color measurements did not seem to have a significant relationship with spectroscopic determination, as parameters a* and b* are located close to each other in the central plane of the principal scatter diagram. Points representing color measurement variables were placed halfway between the anatomical parts and several of the tested phenolic acids and flavonoids. The content of individual compounds and their proportions are hardly determinable using color measurements ( Table 5). The performed Spearman rank correlation showed strong trends between L*, a* and b* parameters, but none of them related significantly to TFC, AA, TAC, chlorophyll and carotenoid content ( Table 5). The correlation matrix shows that there is a low negative relationship between the former and other parameters, except TC, for which the relationship with L*, a* and B* were positive and scarce. For individual phenolic acids and flavonoids, no significant relationship with fruit color was found either.

Genetic relationship analysis
The results of amplification of DNA markers for the five tested Cornelian cherry cultivars showed a large genetic deviation between the analyzed genotypes. Genetic distance values ranged from 0.72 to 0.89  Table 2  ( Figure 3). The greatest genetic distance was observed between cvs. Florianka and Jolico. The lowest distance value was recorded between cvs. P5 and Szafer. Hierarchical cluster analysis tested with the UPGMAM method allowed us to select three clusters. The most similar were cultivars Słowianin, Szafer and P5. Cultivars Jolico and Florianka were placed on separate roots, thus forming two individual clusters.

DISCUSSION
Color is the first determinant of product quality and consumer preferences. Changes in fruit color can be observed not only during storage. Color is also affected by technological processes. [15,16] An increase in fruit lightness may be affected by the decay of anthocyanin colorants, due to the relationship of anthocyanin degradation with the formation of yellow colorants and the increase in yellow color fraction. This may be observed as a decrease in the a* parameter. [5] A different range of the a* parameter (intense red) observed in the study indicates the various content of colorant substances in the tested Cornus mas fruits (Table 2). A similar effect was also observed in cranberry juice stored and preserved in different conditions, for which it was noted that cranberry color brightening may be the result of anthocyanin transformations, which, as the temperature increases (during heating), first becomes more intense due to cleavage of glycosidic bonds, and then the intensity of the color decreases, as a result of anthocyanin oxidation (Roidoung et al. 2016;. Changes in anthocyanin color can also be affected by enzyme activity, i.e., oxidases and the presence of metal ions. The color of anthocyanins in the presence of metal ions depends not solely on what ions occur in the environment, but also on the plant species (e.g., tin ions change the color of strawberries, raspberries and cherries) into pale-red, and the color of black currant into purple-magenta. Meanwhile, the presence of ferric and cupric ions resulted in a brown color . [15,16] The obtained results of TAC are significantly lower than those reported by De Biaggi et al., [17] who stated that the mean value of total anthocyanin content for Cornelian cherry was over 134 mg cyanine glycoside equivalent (CGE)/100 g f.w. In 2011, Kucharska et al. [2] analyzed different Polish cultivars of Cornelian cherry, including Szafer, and observed TAC values of 134.57 mg CGE/100 g and 116.68 mg CGE/100 g, for the first and second cultivar, respectively. The differences between the above results and our results may result from different extraction conditions applied. In the cited studies, acidified methanolic solution was used as an extractant, which significantly improved the migration of ionized molecules of anthocyanins . [2,17] In the study by Popović et al., [18] non-acidified 80% (v/v) ethanol was used. Under these conditions, TAC values ranging from 5.8 to 302.9 mg CGE/100 g fruit were obtained, which are partially consistent with our results. Vitamin C content in the samples ranged from 18.02 to 123.83 mg/100 g f.w (Table 3). In the reference positions, the authors give a wide range of vitamin C content, which is explained by differences in methodology and extraction procedure. Szczepaniak et al. (2019a) noted that differences in AA levels can be significant if we compare data obtained using different analytical techniques, e.g., spectroscopic and titration ones. In the study by De Biaggi et al., [17] the mean AA content amounted to 61.43 mg/100 g f.w., which is similar to that observed in our study for cv. P5. Due to its reducing and antioxidant properties, ascorbic acid is commonly used as a food additive. As an antioxidant, it strongly protects food products against discolouration caused both by enzymatic and nonenzymatic browning. In the latter process, vitamin C reduces the production of ortho-quinone oxidation, keeping the color of the product unchanged . [19] Total carotenoid content ranged from 0.16 to 1.42 mg/100 g f.w. between samples, and no carotenoids were detected in cv. P5 (Table 3). The cultivation method, climatic and soil conditions and, of course, fruit maturity may have a high impact on the content of bioactive compounds (Szczepaniak et al. 2019b). In the presented study, these conditions were unified; therefore, it should be concluded that the diverse composition of the tested phytocompounds is a matter of cultivar only. Comparing the results of colorimetric study, chemical tests and genetic polymorphism analysis, it must be noted that although the first two results of distance projection were similar, the second one was significantly different. Perhaps this variability was affected by other unknown factors or by the presence of interfering compounds not tested in this study. The tested cultivars could be susceptible to environmental factors in different ways, and therefore their metabolism kinetics and products were significantly different. In their study, Ercisli et al. [20] noted that human selection of different traits and natural selection of local conditions, could produce a wide range of Cornus mas genotypes over a long period of time. In another study, [21] several genotypes of the anatomical parts of Cornus mas L. and Cornus officinalis were tested for their mutual relationship and the results of anatomical measurements showed different and worse relationship between the samples than the genetic polymorphism test. Moreover, a large deviation between plants of the same genus originating from the same region was noted by Kalalagh et al. . [22] The color of the fruit is a complex issue. Its parameters differed whether the color of the skin, flesh, skin and flesh together were measured. In several cases, the anatomical part had a stronger statistical power than the cultivar. Cornelian is a rich source of anthocyanins and flavonoids. Its fruits also contain vitamin C and carotenoids. An attempt to statistically quantify the fruit color as a result of the presence of plant secondary metabolites did not succeed.
The results obtained for colorimetric and spectrophotometric assays were not significantly related, but allowed us to construct a statistical model to present differences between the tested cultivars exactly, analogically to the model based on spectroscopic parameters only. The genetic polymorphism study showed different deviations between the tested cultivars than the first two studies.