Molecular characterisation of cornelian cherry (Cornus mas L.) genotypes

Cornelian cherry wild genotypes were used in Bosnia and Herzegovina from three geographically separated regions (Drvar, Mostar and Zenica). Drvar region has average annual air temperatures of 11.0°C and annual precipitation of 1250–1500 mm. Mostar region has average annual air temperatures of 14.6°C and annual precipitation of 1515 mm. Zenica region has average annual air temperatures of 10.1°C and annual precipitation of 782.5 mm. The sample included 20 genotypes of cornelian cherry in Drvar, 20 genotypes in Mostar and 20 genotypes in Zenica. Plant material (leaves) for the DNA extraction was collected from selected plants on each analysed site. Leaves were harvested during spring and summer from each marked tree in 2021. Cold drying of leaf tissue by lyophilisation was performed under vacuum, using lyophiliser (Christ, Model Alpha 1–2 Ldplus). Dried samples were vacuumed in PVC bags and stored at -80°C until DNA isolation. Isolation and genetic characterisation were performed at the Institute of Genetic Engineering and Biotechnology, University of Sarajevo.

DNA was isolated from approximately 50 mg of leaf powder. A modified CTAB protocol was used for DNA isolation (Gasi et al., 2013a, b). After successful DNA isolation from cornelian cherry samples, the PCR protocol was performed. Eight SSR markers were chosen based on their polymorphism reported in previous studies: CM007, CM008, CM010, CM026, CM031, CM037, CM039 and CM043 (Table 1). These markers were originally developed for Cornus mas by Wadl et al. (2014).

Amplification of microsatellite sequences was performed in a PCR device ABI GeneAmp® PCR System 9700 (Applied Biosystems, Foster City, CA, USA). Fluorescently labelled primers were used for amplification in order to be able to multiplex and analyse the PCR product on a DNA genetic analyser. Amplification of selected loci was performed in two separate PCR reactions (mix 1 and mix 2) with four microsatellite loci each (Table 2). All PCR reactions were conducted in a total volume of 15 μL, containing 2.5 mM of MgCl₂, 1 × PCR buffer, 0.2 mM dNTPs, 0.2 U · μL^-1 of TrueStart Taq polymerase (Thermo Scientific) and 10–50 ng of template DNA. The temperature conditions for the amplification reaction were the same for both PCR reactions (Table 3). Before amplification, DNA was tested for quality by horizontal electrophoresis on agarose gel.

Allele sizes were determined by analysis of PCR products on an ABI 3500 genetic analyser Applied Biosystems), by vertical capillary electrophoresis. Aliquots of amplified DNA (1 μL), mixed with formamide and GeneScan-350 ROX Size Standard (Applied Biosystems) were run on an ABI 3500 genetic analyser. LIZ 500 (Applied Biosystem) was used as an internal standard. SSR profiles were scored using the GeneMapper ID 5 Software (Applied Biosystems).

Allele frequencies, private alleles, detected and effective number of alleles, observed and expected heterozygosity, as well as within-group inbreeding (Gasi et al., 2013b; Wadl et al., 2014) and deviation from Hardy–Weinberg (HW) equilibrium were calculated in GenAlex v6.503 (Doyle and Doyle, 1987). The ratio of the effective and detected number of alleles as well as its statistical significance were calculated using the ALRATIO script (Wadl et al., 2014). Fst genetic differentiation analysis (Cullings, 1992) and analysis of molecular variance (AMOVA) between observed groups was performed in GenAlex v6.503 software (Nei, 1987). Genetic distance analysis according to the Nei (1973) and principal coordinate analysis (PCoA) analysis, based on the results of individual and population genetic distance, were performed in the same software. Based on the results of the interindividual genetic distance, a neighbour-joining (NJ) dendrogram was constructed in the MEGA X software (Peakall and Smouse, 2012).

Genetic analysis of cornelian cherry involved the use of eight microsatellite markers for genetic characterisation of 60 examined genotypes and are given in Table 4. The markers, which successfully amplified PCR product, were found to be highly polymorphic (Table 4). A total of 25 private alleles were detected in these three populations. There are 7 private alleles in the Drvar population, 14 in the Mostar population and 4 private alleles in the Zenica population. Private alleles are alleles that are found only in a single population among a broader collection of populations. These alleles have proven to be informative for diverse types of population-genetic studies. Different processes can lead to the appearance of private alleles (isolation, inbreeding and self-fertilisation in wild populations, or targeted selection during cultivar breeding). A large number of private alleles in these three populations show that the genetic material of cornelian cherry from Bosnia and Herzegovina is unused for breeding purposes. The Mostar population stands out in particular (14 private alleles). The number of alleles in Drvar group was 8.6 per locus, in the Mostar group 9.8, in the Zenica group 7.6 and for all groups 8.7. The total average number of detected alleles in all groups was 8.70 (Table 4).

Parameters of heterozygosity are given in Table 4. The number of detected alleles (Na) within the Drvar population shows that it is the highest for the CM031 locus, and the lowest for the CM043 locus. For the Mostar population, the largest number of alleles was observed at the CM010 locus, and the smallest at the CM031 and CM043 loci.

When it comes to the Zenica population, the lowest number of alleles was recorded at the CM043 locus, and the highest at the CM031 locus. The average number of detected alleles in the Drvar group was at eight SSR loci.

On the contrary, the effective number of alleles (Ne) within the Drvar population was recorded at the CM007 locus, and the smallest at the CM026 locus. For the Mostar population, the highest effective number of alleles is observed at the CM010 locus, and the lowest at the CM026 locus. The number of effective alleles within the Zenica population shows that it is the highest at the CM007 locus, and the lowest at the CM043 locus. The total average effective number of alleles in all groups was 5.48 per locus.

In a study conducted on cornelian cherry by Wadl et al. (2014), the number of alleles per locus ranged from 1 to 11 and averaged 6.73, whereas the effective number of alleles per locus ranged from 1.00 to 5.53 with an average of 3.29. Another study on 447 wild grown cornelian cherry genotypes represented in Austria, using seven microsatellite markers, came to the following results: the average number of alleles per locus was 10.57 and the effective numbers of alleles was 3.86 (Borroto Fernandez et al., 2023).

From Table 4, it can be concluded that the highest observed heterozygosity (Ho) in the Drvar population was recorded for locus CM007, and the lowest was for CM008 and CM010. Also, the highest expected heterozygosity (He) in the mentioned population was observed for locus CM007, and the lowest for CM026. Within the Mostar population, the highest observed heterozygosity was recorded for the CM007 locus, and the lowest for the CM043 locus. On the contrary, the highest expected heterozygosity was observed at the CM010 locus, and the lowest at the CM026 locus. In the Zenica population, the highest observed heterozygosity was recorded at the CM007 locus, and the lowest at the CM043 locus. For the expected heterozygosity parameter within the mentioned population, the highest value was observed at the CM007 locus, and the lowest at the CM043 locus. In all three populations, the mentioned loci are 100% polymorphic.

Previously, Peakall and Smouse (2012), Wei and Becher (1984) and Kumar et al. (2018) reported the observed heterozygosity from 0.00 to 0.71, while the expected heterozygosity spanned from 0.00 to 0.82.

The examination of the relationship between the effective and detected number of alleles (R) revealed no statistically significant reduction in the effective number of alleles compared with the detected number (p(Ra) > 0.05). This finding is highly significant as it indicates that the majority of detected alleles at the observed locus contribute significantly to the genetic diversity within the observed groups (Table 4).

Analysing the inbreeding coefficient (F) for the Drvar population showed the highest value at locus CM010 and the lowest at locus CM007, indicating signs of outbreeding. In the Mostar population, the highest F was observed at locus CM043, while the smallest was at locus CM007, once again suggesting outbreeding. Concerning the Zenica population, the highest F was recorded at the CM026 locus, and the lowest at the CM039 locus, where a negative value implies outbreeding. The total inbreeding across all three observed populations was calculated as 0.056 (Table 4).

The results of the test for deviations from HW equilibrium using the χ² test are given in Table 5. Within the population of Drvar, four loci deviate statistically significantly from HW equilibrium, two in the population of Mostar, and one locus in the population of Zenica. From all of the above, it can be concluded that the observed cornelian cherry groups show a moderate level of heterozygosity without significant inbreeding, with a good basis when it comes to allelic diversity.

However, certain loci (CM008, CM010, CM031), especially within the Drvar population, show a tendency towards a more intense level of inbreeding, which may indicate an upcoming trend of heterozygosity reduction. Therefore, it is necessary to analyse possible changes in genetic heterozygosity in certain annual examinations, all with the aim of preventing its rapid reduction, which would ultimately lead to inbreeding depression.

The results of the PCoA analysis are shown in Figure 1. There is a tendency of differentiation of the samples from the three mentioned populations, but a clear differentiation of the samples was not observed. This is consistent with the results of genetic differentiation (Fst, pFst and AMOVA).

Figure 1.

The total genetic differentiation is also confirmed by the AMOVA, which indicates that 3% of the total genetic variation is between groups, 8% between individuals and 89% within individuals (Table 6).

The highest genetic differentiation (Fst) between the observed groups was recorded for the CM043 locus, while the smallest was for the CM007 locus (Table 7). The total genetic differentiation is 0.042, which is also confirmed by the AMOVA, which indicates that 3% of the total genetic variation is variation between groups, 8% between individuals and 89% within individuals (Table 6). In a study of Fst was between 0.05 and 0.15.

The results of pairwise Fst analysis (pFst) indicate that the greatest differentiation is between the populations of Mostar and Zenica (0.037), between Drvar and Zenica (0.025) and between Drvar and Mostar; the smallest is 0.025 (Table 7). The results of the genetic distance according to Nei confirm the mentioned results and point to the fact that the biggest difference was observed between the populations of Mostar and Zenica (0.316), then between Drvar and Zenica (0.285) and the smallest between Drvar and Mostar (0.251).

The present research revealed significant differences between individuals and within population groups. For a given SSR, high genetic variation in a sample of population might be due to an increase in gene flow or the mutation of a number of repeats of a given genotype. Additionally, this high genetic variation could be due to the natural selection process within the studied genotypes of cornelian cherry. The lower genetic variation between these populations may be related to the possibility of sharing a common ancestry, despite growing in different regions.

The constructed NJ dendrogram based on the results of the interindividual genetic distance analysis is shown in Figure 2. The NJ dendrogram indicates similar results of the PCoA analysis where there is a certain grouping of samples, but no pronounced grouping of complete populations can be observed, which would indicate a more pronounced genetic differentiation.

Figure 2.