Phenotypic Characterization of a Wild-Type Population of Cornelian Cherries (Cornus mas L.) from Austria

Cornelian cherry (Cornus mas L.) belongs to a group of fruit and nut species growing in Europe considered to be underused economically, although it has been recognized as a potential regional “superfood” and as a source of valuable bioactive compounds. Phenotyping fruits of 30 accessions of an Austrian wildtype population of C. mas allowed to evaluate their nutraceutical potential. Ten fruits per accession were characterized by morphological and morphometric approaches. Biochemical analyses were performed to determine the respective amounts of vitamin C, sugars, anthocyanins, iridoids and flavonoids. Both datasets were subjected to statistical analyses. Morphological and morphometric characterization and biochemical analyses enabled the identification of the individuals with the highest economic value. Statistical treatment of data identified the most significant principal components. The first phenotypic profiling of bioactive compounds of wildtype C. mas in Austria yielded a high variability. Dealing with wildtype plants, this is not surprising. However, our results allow to select among the Austrian C. mas accessions the most interesting individuals for further breeding of this alternative fruit with interesting nutritional values.


Introduction
The genus Cornus belonging to the family Cornaceae comprises about 65 species, characterized by colourful and attractive flowers and fruits (Yilmaz et al. 2009). Most Cornus species are used as ornamentals, such as Cornus florida or Cornus kousa, while only few species are grown for their fruits, including Cornus mas and Cornus officinalis (Ercisli et al. 2011, Czerwińska andMelzig 2018).
Cornus mas L. (Cornelian cherry, European cornel or dogwood) is predominantly distributed over Southern and Central Europe, the Black Sea basin and the Caucasus (Da Ronch et al. 2016;Schramayr 2009).
Cornus mas has been used as food and as a medicinal plant for millennia in Eurasia (Gismondi et al. 2012), and all parts of the plant are valuable (Dinda et al. 2016; Czer- Margit Laimer margit.laimer@boku.ac.at 1 Plant Biotechnology Unit, Institute of Molecular Biotechnology, Department of Biotechnology, University of Natural Resources and Life Sciences Vienna, Vienna, Austria wińska and Melzig 2018). Cornelian cherry fruit stones moreover may have had a symbolic role in funerary rituals by the Mesolithic community (Filipović et al. 2020). Their fruits are a rich source of vitamin C, polyphenols, organic acids, sugars, and minerals (Seeram et al. 2002;Tural and Koca 2008;Rop et al. 2010;Hassanpour et al. 2011) that might contribute to a healthy nutrition and be considered as a "superfood" (Bayram and Ozturkcan 2020).
In fruits, the amounts of soluble sugars, including fructose and glucose, not only play an essential role as a source of energy, structure and functionality of the whole plant (Ma et al. 2014), but are also essential for the growth and quality of the fruits. The sugar content in Cornus mas plays an important role in consumer satisfaction and consumption (Bayram and Ozturkcan 2020). Cornelian cherry fruits are consumed fresh or dried, are used to produce jam, stewed fruit, paste, marmalade, dried fruit pulp and syrup, several types of soft drinks and used for medicinal purposes too (Tural andKoca 2008, Turhan et al. 2007). Fruit extracts are also used in Europe for cosmetic purposes, replacing synthetic astringent substances, and are claimed to exert a favourable action on the human complexion (Polinicencu et al. 1980).
In Austria, Cornelian cherry is the leading species in the Pielach Valley in Lower Austria. As a result of a wide range of products manufactured in the zone, such as jams, drinks, juices, ice creams and vinegars, this area is marketed as a "Valley of Cornelian Cherries" for tourism (Schramayr 2009). Cornelian cherries in Austria have not been systematically studied although their healthy ingredients make them a potential regional superfood (Borroto Fernandez and Laimer 2021).
Here we describe our effort to select the most interesting individuals among a population of over 400 centenary plants growing wild in the region of the Pielach Valley and two neighbouring valleys (Lower Austria). This was achieved by characterizing both morphological traits, mainly linked to the use as food (such as appearance and taste), and most secondary metabolites such as flavonoids, iridoids and cornusides, which are mainly interesting for health-related effects.
The aim of this study was to provide information about the variability of morphological traits as well as nutrients and pharmaceutical ingredients in 30 Cornus mas Austrian accessions. This phenotypic characterization might explore Cornelian cherry accessions as a source of valuable material for further selection in Austria.

Plant Material
The entire population of cornelian cherry (Cornus mas L.) accessions from the Mostviertel region in Lower Austria comprised more than 400 individuals from 40 sites at different altitudes (280-600 m above sea level), in open and closed forest areas [Borroto and Laimer, unpublished data]. A total of 30 Cornelian cherry (Cornus mas L.) accessions, among 152 individuals, were selected from 18 sites in Pielach Valley (P) and Gölsen Valley (G; Table 1). The selection was made based on distinguished attractivity and famers' preferences. Plants of different age and with fruits of heterogeneous sizes, shapes and colours were selected (Table 1). According to the age, they were two plants of about 10 years old, one plant older than 50 years, 11 plants older than 100 years, 15 plants older than 200 years, and the oldest plant was older than 400 years. We selected plants with fruits of different sizes, shapes and colours (Table 1). One plant had yellow fruits, which seems a rare mutation, because we only encountered three plants with this phenotype in more than 400 accessions studied. Three plants presented orange fruits, nine had light red fruits, eight medium red and nine dark red fruits. Approximately 100 g of fruits from these accessions were picked at the stage of full maturity.

Morphological Characteristics
Each sample for the morphological analysis consisted of 10 fully matured fruits and they were evaluated for their morphological characteristics (Table 2). Stone weight and fruit weight were measured using a digital bench balance with a sensitivity of 0.001 g. Fruit length (mm), fruit width (mm), stone width (mm) and stone length (mm) were measured by a digital calliper with a resolution of 0.01 mm (Walter ® , Salzburg, Austria). Fruit flesh ratio was counted considering fruit and seed weight; flesh/seed ratio was calculated by mean fruit weight-mean seed weight/mean seed weight.
Vitamin C in an exploratory analysis was measured on a reduced number of accessions by HPLC-UV and electrochemical detection after ascorbic acid was extracted from the sample matrix with an acetonitrile-based solu-tion and isoascorbic acid was used as an internal standard (AGROLAB, Kiel, Germany.).

Statistical Analysis
Data recording for the morphological characteristics and chemical composition was performed in Microsoft Excel 2019 (Microsoft, Redmont, WA, USA). Morphological data were subjected to ANOVA and a post hoc Tukey's HSD test to separate and compare means if significant differences (p < 0.05) were detected. Physical properties were calculated using Pearson's correlation coefficient. Mean, maximum, minimum, and standard deviation (SD) values for the chemical characteristics in the study were calculated ( Table 3). The morphological and chemical characteristics (except vitamin C) were analysed by principal component analysis (PCA). The results of the PCA were used to construct two-dimensional scatter plots of the first two factors. Statistical analyses were performed using the XLSTAT statistical software program (Addinsoft Inc., New York, NY, USA).

Morphological Characteristics
The present study found statistical differences in all morphological traits among the 30 accessions of Cornelian cherry analysed (p < 0.05). Table 2 shows mean values of the morphological characteristics determined in 30 Austrian accessions.
The average fruit weight of Cornelian cherry in this study ranged from 1.38 g to 2.58 g, which is within the range of 1-10 g of previous studies observed in Iran (Hassanpour et al. 2012), Serbia (Bijelić et al. 2011;Mratinić et al. 2015), Montenegro (Jaćimović et al. 2015), Poland (Szot et al. 2019) and Turkey (Yilmaz et al. 2009), although a Serbian genotype with 14.55 g of fruit weight was described (Bošnjaković et al. 2012). An important question associated with the dimension of the Serbian genotype can be attributed to the fact that the genotypes under study were from secondary populations (Bošnjaković et al. 2012). The fruit flesh weight was between 1.1 g and 2.36 g; however, this value was reported to be higher in Iran (Hassanpour et al. 2012).
The length and width of the fruits in the selected Austrian accessions varied from 13.46 mm to 19.08 mm and 9.12 mm to 13.77 mm, respectively, which is consistent with the results of other studies where fruit lengths are between 10 and 23 mm (Bayram and Ozturkcan 2020) and fruit width between 7.43 and 23.51 mm depending on the region (Yilmaz et al. 2009, Bijelić et al. 2011Hassanpour et al. 2012).
Interestingly, the fruit length/width index and the stone length/width index, correspond exactly to the fruit and stone shape, respectively. Thus, the described fruit shape (round or elliptical) could be objectively confirmed and calculated with this index.
Obviously, the farmers prefer plants with large fruits of different colours, which make the harvest easier. However, there is no way to correctly judge the yield from the exterior appearance (length and width) and the weight of the fruits, since it is the proportion of the kernels that finally determines the fruit flesh percentage that can be obtained from the crop. Instead, we would like to argue that the last column in Table 2 (% fruit flesh) contains the value of most interest for the farmers. Nevertheless, one should keep the caveat in mind that this might vary in this rain-fed crop with the availability of annual precipitations.
Among the plants analysed, two were in the range of 76.81-79.99%. Usually they would be discarded, unless they had another interesting trait, which for plant 6 is clearly the yellow fruit colour. Usually, plants with yellow fruit are considered less vigorous (Pirc, personal communication); however, the administration of yellow fruit extracts of Cornus mas increased the level of reduced glutathione and mean cell haemoglobin in diabetic rats (Dzydzan et al. 2019).
In the medium range, which encompassed the majority of the accessions, we found ten plants with 85.03-85.86% fruit flesh and six plants with 86.03-86.90% fruit flesh. In the high range we found two plants with 87.16-87.75% fruit flesh, two plants with 88.15-88.20% fruit flesh and a single plant with 91.72% fruit flesh.
Plant 29 had large elliptical light red fruits with an attractive appearance and the heaviest fruits with the highest flesh weight. This plant also had the highest percentage of fruit flesh. Plant 6 was the plant with the lightest and smallest fruits, and the second lowest percentage of fruit flesh with medium-sized elliptical fruits, but of a yellow colour. Plant 12 with the longest and second widest fruits had large elliptical fruits of an orange colour. The fruits of plant 23, which were catalogued as medium round and medium red, possessed the heaviest stone weight; in fact, plant 23 had the lowest fruit flesh percentage.
Plant 16 had the roundest fruits and the second highest fruit weight and a high fruit flesh percent. Plant 4 had medium-sized, round, dark fruits with the narrowest stone width and the lowest stone weight, and a high percentage of fruit flesh.
Interestingly, plants 21 and 22, which stood close to 32, represent two extremes: Plant 21 had the longest kernels and the highest stone length/width index, medium elliptical size, medium red fruits with a lower percentage of fruit flesh; while plant 22 had the shortest stones with a medium percentage of fruit flesh, medium size, round and medium red fruits. Table 3 shows the Pearson correlation matrix for the morphological traits in this study with a significant positive correlation. There was a significant positive correlation between fruit weight, fruit length, fruit width and flesh weight. Stone weight is positively correlated with stone width. The flesh/stone ratio is positively correlated with fruit flesh and negatively correlated with stone width. Fruit length is positively correlated with the stone length. Fruit width is negatively correlated with the index of fruit length/fruit width. The index of fruit length/fruit width is positively correlated with the index of stone length/stone width. The stone width is negatively correlated with the index of stone length/stone width as well as with the percentage of fruit flesh.
The observed correlation between fruit weight and fruit width would also allow us to sort Cornus mas in Austria according to the fruit width, contrary to what was observed by Hassanpour et al. (2012). Instead, we would like to argue that the percentage fruit flesh (last column in Table 2) contains the value of most interest for the farmers.

Bioactive Compounds
It has been claimed that Cornelian cherry fruits contain twice as much ascorbic acid (vitamin C) as oranges (Seeram et al. 2002); however, it could fluctuate between 11 and 360 mg/100 g (Bayram and Ozturkcan 2020). Although the vitamin C values from Pielach Valley recorded in this study are lower than other genotypes examined in Iran (240-360 mg/100 g; Hassanpour et al. 2011Hassanpour et al. , 2012, the differences in the values of vitamin C could depend on the cultivar, cultivation condition, genotype as well as the methods used for content determination (Bayram and Ozturkcan 2020).
The vitamin C content for all Cornelian cherry analysed varied from 5.08 to 92.10 mg/100 g with an average of 29.82 mg/100 g. Our mean value is close to the value re-ported in Serbia and Montenegro, where it ranges between 22.44 and 103 mg/100 g (Bijelić et al. 2011;Jaćimović et al. 2015;Ognjanov et al. 2009), and it also resembles values recorded in Greece (14-103 mg/100 g; Pantelidis et al. 2007) and Turkey (11-106 mg/100 g; Aslantas et al. 2007;Copur et al. 2003;Pirlak et al. 2003).
The carbohydrate compounds measured in the 30 Austrian Cornelian cherry accessions were glucose and fructose with values ranging between 4.54 and 7.84 g/100 g and 2.86 and 5.44 g/100 g, respectively. These results are in agreement with results reported for Bulgaria (Petkova and Ognyanov 2018), Turkey (Okan et al. 2019), Russia (Perova et al. 2014, Poland (Antolak et al. 2017) and Montenegro (Martinović and Cavoski 2020) that oscillate between 1.90 and 3.8 g/100 g for fructose and between 2.5 and 12.26 g/100 g for glucose. Contrarily, the glucose and fructose values detected in this research are significantly lower than those from the Gemer Region of Slovakia (Brindza et al. 2009).
Glucose and fructose in Cornelian cherry have a crucial impact on producing drinks, sweets, syrup, marmalade as well as fresh fruit consumption because they determine the palatability of the fruits (Tural and Koca 2008;Turhan et al. 2007) but they might also influence-as for black carrot, elderberry and strawberry-the stability of anthocyanins (Sadilova et al. 2009).
The major anthocyanin compounds reported in Cornus mas are Cy 3-Gal and the Pg 3-Gal (Dinda et al. 2016). Our analyses revealed that cyanidin Cy 3-Gal values range between 0.063 and 2.886 mg/g and Pg 3-Gal between 0.064 and 2.355 mg/g. These findings are in line with previous studies published by other researchers Pawlowska et al. 2010;Moldovan et al. 2016;Sozanski et al. 2014;Dzydzan et al. 2019). Although our maximum values are slightly higher than the aforementioned studies, they are lower than in the variety 'Raciborski' in Poland . Alongside anthocyanins, Cornelian cherries also contain the iridoids loganic acid and cornusides (Dinda et al. 2016;Szczepaniak et al. 2019).
Iridoids are documented to be involved in plant defences against insects (Dyer et al. 1996;Richards et al. 2012), and iridoids isolated from Cornus species possess anti-inflammatory and antioxidative properties (Czerwińska and Melzig 2018). The mean content for loganic acid and cornuside in our investigation was 12.364 mg/g and 0.709 mg/g, respectively, which places the Austrian accessions investigated closer to the range noted for some Polish varieties Szczepaniak et al. 2019, Sozanski et al. 2014.
Flavonoids are the second largest constituents in fruits and major constituents of leaves of Cornelian cherry (Dinda et al. 2016). In our fruit analyses, Q 3-Glr content ranged between 0.115 and 0.331 mg/g, which is considerably lower than data reported from other regions (Pawlowska et al. 2010;Moldovan et al. 2016;Dzydzan et al. 2019). However, the values resemble data obtained in Serbia (4.45-151.82 µg/g) by Bajić-Ljubičić et al. 2018 and in Montenegro (13.40-29.66 mg/100) by Martinović and Cavoski (2020). The differences observed could be attributed primarily to the genotype of the plants as well as to the geographical location and the prevailing climatic conditions.
Quercetin has been classified as an antioxidant (Satyanarayana et al. 2001) and has been shown to prevent the nephrotoxic effect of cisplatin, without affecting its anti-tumour activity (Sanchez-Gonzalez et al. 2011). This flavonoid is also responsible for preserving hippocampal neurons and enhancing the synaptic plasticity (Akang et al. 2019).
Determining some principal bioactive compounds in Cornus mas accessions in Austria enabled the evaluation of their profile and compare it with those already described K  in other regions. This allowed us to evaluate the potential of the Austrian accessions as functional food as well as the importance of these results as part of breeding and selection programmes.

Principal Component Analysis
Principal component analysis produced 17 principal components (PCs) with eigenvalues greater than 0.6 (PC7), explaining 92.87% of the total variability observed (Fig. 1, shows 11 out of the 17 principal components; Table 4). Absolute values greater than 0.5 were found in the first seven most interesting PCs. Based on those criteria, the first seven PCs were taken into consideration (Table 5).
Principal component analyses showed that the first seven principal components (PC1-PC7) explain 25.9, 49.6, 64. 8, 75.4, 83.1, 89 and 92.87% of the variation, respectively (Table 5). Bold values indicate correlation coefficients with value greater than or equal to 0.5 in absolute value In the first component (PC1) the characteristics with higher scores are related to fruit weight, length and width as well as to flesh weight and account for 25.88% of the variability. This first component might be regarded as corresponding to the fruit size.
The second factor (PC2) correlated mainly with stone width and stone weight, flesh/stone ratio and % fruit flesh (r = -0.860, -0.851, 0.874 and 0.926, respectively) and to a lesser extent with cornuside (r = -0.499) and has a contribution of 23.75% of the total variance. Therefore, it might be called the stone dimensions, fruit flesh and cornuside component.
Factor PC3 represents 15.11% variability associated with stone and fruit shape (r = 0.853) calculated from index fruit length/fruit width and index stone length/stone width along with stone length, which might be analysed as the fruit shape component.
Factor PC4 with 10.65% of total variability is a factor correlated with cyanidin 3-O-galactoside and pelargonidin-3-O-galactoside could be called an anthocyanin component.
Factors PC5, PC6 and PC7 represent 7.72, 5.9 and 3.85% of the total variance, respectively. The highest correlation for PC5 is loganic acid and fructose while for PC6 it is quercetin-3-O-glucuronide (r = 0.598) and for PC7 it is glucose.
These principal components analyses indicate a substantial morphological variation between the accessions, which results in the decisive importance of morphological characterization in the evaluation of genetic diversity. Among the bioactive compounds considered the three most influential were cornuside, cyanidin 3-O-galactoside and pelargonidin-3-O-galactoside, as shown in PC2 and PC4. At this stage of understanding we believe that although the correlation between all groups of traits is not the same, all are represented in the first seven PCs, suggesting that these traits could be genetically related.
The accessions distribution on PC1 (fruit size) and PC2 (the component reflecting stone dimensions, fruit flesh and cornuside) are plotted in Fig. 2. The majority of accessions are located in the central part of the scatter plot defined by the oval. However, it is possible to assign accessions into main groups. The accessions located in the triangle share a high fruit weight and high fruit length values. In addition, those accessions shared values ranging from high to medium by flesh weight and fruit width and the fruits have light red colour. Genotype 29 belonging to this group had the highest value for fruit flesh as well as for the flesh/ stone ratio but had lower average cornuside values. A third group (rectangle) contains accessions that were chosen in this exploration for the light red colour (Laimer and Borroto, unpublished data) and at the same time they shared high fruit flesh (%) values.
In the lower part (23 and 6) and top left (4) of the figure, three outliners are found in which accession 6 had the lowest values for fruit weight, flesh weight and fruit width and together with 23 had the lowest values for flesh/stone ratio and fruit flesh. However, 23 had the highest cornuside, stone weight and stone width values, while 4 presented the lowest value in stone weight and low values for fruit weight and flesh weight, although the percentage fruit flesh is considered one of the highest in our study. As an interesting aside, accession 4 and 23 had red dark fruits, whereas 6 is a yellow genotype.
Based on the results in the present study, the observed accessions could be categorized as a representative natural population of Cornus mas in Lower Austria with a diverse morphological characteristic and a valuable source of K bioactive compounds; this is the first time that a phenotypic profile of Austrian Cornelian cherry fruits has been accomplished. As a result, a representative number of the accessions analysed were selected  for the first in vitro Cornelian cherry germplasm collection in Austria due to its importance for the conservation of natural populations as well as for future breeding programmes.

Conclusion
This study highlights that Cornus mas accessions from Lower Austria, still neglected and underutilized, represent a natural population with diverse morphological characteristics and with potential as a beneficial source of bioactive compounds. It is the first time a phenotypic profile of Austrian Cornelian cherry fruits has been completed. Our findings will enable the selection of interesting genotypes in the development of a breeding strategy for genotypes adapted to future needs for a sustainable production in the region.