Plant height variation and genetic diversity between Prunus ledebouriana (Schlecht.) YY Yao and Prunus tenella Batsch based on using SSR markers in East Kazakhstan

Study of genetic structure using SSR

This study investigated three isolated populations of P. ledebouriana from two mountain geographic ranges (Altai and Tarbagatai) and one P. tenella from Eastern Kazakhstan. Among these four populations, the first one (1-UR) was collected on several isolated gorges in the state national natural park “Tarbagatai” at the height of the shrub belt. The materials of the second and third populations were collected in the cold and xerophytic mountain regions of the Kalba (2-KO) and Narym (3-KA) ridges. The population of P. tenella (4-UK) was collected from a small hilly plain (steppe) zone in the outlines of the Kalbinskiy and Ulbinskiy ridges in the border zone of the city of Ust-Kamenogorsk and the village of Novo-Akhmirovo (Table S1). The plant height of P. ledebouriana was measured according to Goloskokov (1972). In total, 20 leaves from each of the 60 P. ledebouriana and 20 P. tenella plant populations were collected. The distances between populations were at least 100 kilometers, and plants within a population were selected at a distance of at least 50 m from each other.

Fresh plant leaves were used for DNA extraction. Total DNA was isolated from crushed leaf powder according to the Cetyl trimethyl ammonium bromide (CTAB) protocol with double purification with chloroform (Doyle & Doyle, 1987). The quality and concentration of DNA were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and electrophoresis in 1% agarose gel. DNA concentration was normalized to the working concentration for further analysis.

Twenty-two SSR markers of the nuclear genome were used as DNA markers and selected according to Mnejja et al. (2005). PCR amplification was performed in 10 ul reaction volume containing 20 ng template DNA, one PCR buffer, 1.5 mM of MgCl2, 0.2 mM of dNTPs, 0.4 mM of each primer, and one unit of Taq DNA polymerase (Sileks, Badenweiler, Germany). PCR conditions were set at 95 °C for 1 min, followed by 35 cycles of 94 °C for 30 s, 50–65 °C for 30 s, 72 °C for 30 s, and 5 min at 72 °C for final elongation. PCR products were separated on a 6% polyacrylamide gel using 0.5x Tris-borate-EDTA buffer. DNA fragments were identified using an ethidium bromide staining procedure. Alleles were determined using 100-bp ladders (Thermo Fisher Scientific). Visualization was performed using the GelDoc XR+ gel documentation system (Bio-Rad, Hercules, CA, USA).

Statistical analysis and polymorphism information content

Descriptive statistics and t-tests (SPSS Statistics v.27.0; https://www.ibm.com/products/spss-statistics) were used to describe morphological features and identify population differences (plant height across different populations of P. ledebouriana and P. tenella).

Each DNA fragment obtained was treated as a separate character and evaluated as a discrete variable. Accordingly, rectangular binary data matrices were obtained for SSR markers.

To assess the effectiveness of markers, the following polymorphism indices were used: marker index (MI), resolving power (Rp), observed heterozygosity (HO), expected heterozygosity (HE), polymorphism information content (PIC values), inbreeding coefficient (FIS), and fixation index, (Fst) as estimated by Weir & Cockerham (1984).

The genetic diversity of P. ledebouriana, including the Shannon diversity index (I), was assessed using PopGene (Yeh et al., 2000). Wright’s F-statistics was calculated for each SSR locus using the GenAlEx 6.5 program (Peakall & Smouse, 2006). Analysis of molecular variance (AMOVA) was performed across organizations using GenAlEx 6.5 (Peakall & Smouse, 2006). Principal coordinate analysis (PCA) was performed using the Numerical Taxonomy and Multivariate Analysis System (NTSYS-pc) program (Rohlf, 1998). The STRUCTURE program also applied Bayesian cluster analysis (Pritchard, Stephens & Donnelly, 2000). The dendrogram was constructed using the PAST program and unweighted pair group method with arithmetic mean (UPGMA) algorithm and Boot N:1,000 (Joshi et al., 2000). Cluster analysis of P. ledebouriana and P. tenella was studied using STRUCTURE. An admixture model was used, which made it possible to analyze the frequencies of admixtures and correlated alleles. Five independent simulations were run, each including 100,000 burn-in steps and a subsequent 100,000 Markov chain Monte Carlo (MCMC) iterations.

Variability of plant height among populations

The samples of three P. ledebouriana and one population of P. tenella were assessed according to plant height (Fig. 1). The most significant plant height was recorded for the 3-KA population (2.09 ± 0.06 m), which was the mountain population’s highest elevation above sea level (Table S1). Similar parameters for plant height were recorded in representatives from two other mountain populations: 2-KO and 1-UR, 1.79 ± 0.03 m and 1.78 ± 0.03 m, respectively. The lowest height values of plants in the steppe populations of P. tenella were recorded in 4-UK (1.41 ± 1.04 m). The T-test confirmed a statistically significant difference between the plant height of the three populations of P. ledebouriana: 3-KA, 2-KO, and 1-UR (P value = 2.3e−15). The t-test suggested that all three mountainous populations had significantly higher plants than the steppe population P. tenella 4-UK (P < 0.0001).

Allelic variation of SSRs

The assessment of allelic variations in P. ledebouriana and P. tenella showed that 19 out of the studied 22 markers were polymorphic (Table S2). Figure 2 shows two gel snapshots of different populations with high polymorphism. It appeared that CPDCT038 had three loci. Information on these 19 polymorphic SSR loci, including the sizes of bands, is presented in Table 1.

Results showed that the three SSR loci had two alleles, 11 SSR loci had three, three SSR markers had four, and two loci had five. The results of genotyping using 19 SSR loci and their statistical details are shown in Table 2. It was determined that 19 loci, according to the polymorphism level (PIC values) could be separated into three subpopulations. The first group was comprised of five SSR loci with a PIC above 0.5, the second group of eight loci with a PIC between 0.3 and 0.5, and the third group of the remaining six loci with a PIC below 0.3. The mean PIC value was 0.38 (Table 3).

The genetic diversity assessment using Nei’s index suggested that the highest genetic diversity was in P. tenella 4-UK (0.622), followed by three populations of P. ledebouriana: 1-UR, 2-KO, and 3-KA (Table 2). The average genetic diversity index for the three populations of P. ledebouriana was 0.501. To compare inter-population diversity, two main scenarios were used. The first included the assumption of the unity of four populations as representatives of the species P. ledebouriana, and the second scenario assumed the presence of three populations of P. ledebouriana and one population of P. tenella. The AMOVA suggested that the total genetic diversity in P. ledebouriana could be partitioned as 73% within populations and 27% between populations. The evaluation of the partitioning of the genetic variation across four populations (including 4-UK P. tenella) resulted in a decrease in the level of variation within populations (63%) and an increase in variation between populations (37%).

The t-test was applied to test associations between SSR markers and plant height in samples of both species. The results showed that nine SSR loci were statistically associated with plant height (Table 4).

Population structure in samples of four studied populations using SSR markers

All P. ledebouriana and P. tenella samples were analyzed for population structure using the package STRUCTURE based on genotyping data from 19 polymorphic SSR markers. The structure was assessed using results from K = 2 to K = 10. The assessment of K plots suggested that starting from K = 3 and K = 4 (Fig. 3), plants from population 4-UK were separated from three populations of P. ledebouriana. Interestingly, the population 3-KA, growing at the highest elevation and characterized by the lowest level of genetic diversity within four studied populations (Table 2), was drifting apart from other groups of plants in step K = 2. In the analysis of principal coordinates (PCoA), it was determined that the first and second principal coordinates described 49.05% and 41.16% of the variability, respectively (Fig. 4). PC1 effectively separated 3-KA from the other three populations. At the same time, PC2 allowed for the differentiation of 4-UK from 1-UR and 2-KO. In addition, the UPGMA dendrogram was built based on the genotyping results for samples in four populations. The results suggested that population 4-UK formed a distinct cluster, and only one sample from that population (4-UK_07) was positioned close to the cluster with the domination of samples from population 3-KA. Likewise, in PCoA analysis, the UPGMA dendrogram distinguished 3-KA from 1-UR and 2-KO. At the same time, the latter populations had a mix of samples in several clades (Fig. 5).

The phylogenetic relationship between P. tenella and P. ledebouriana

The phylogeny of the genus Prunus is complex. Previous reports provided contradictory results on the relationships of plum species (Badenes & Parfitt, 1995). Nevertheless, it has been well established that P. tenella belongs to the section Amygdalopsis within the subgenus Amygdalus (Avdeev, 2016). The complexity of taxonomy in species within this section could be more precise, as there are questions about the relationship among different taxa. These questions include the taxonomic relationship between P. tenella and P. ledebouriana, where the former is widespread in the Eurasian continent, and the latter is limited to mountainous populations of East Kazakhstan, particularly in the Altai mountains (Zhukovsky, 1971; Dzhangaliev, Salova & Turekhanova, 2003; Myrzagalieva et al., 2015; Myrzagalieva & Orazov, 2018). Several reports proposed that these two species have minor differences and could be considered one species (Sokolov, Svyazeva & Kubli, 1980; Qiu et al., 2012).

Conversely, publications suggest that P. ledebouriana and P. tenella have sufficient morphological differences to separate them into two distinct species (Zaurov et al., 2015; Orazov et al., 2020). Plant height is one of the most prominent traits used to differentiate these species (Mushegyan, 1962; Vintereoller, 1976). In this work, all plants from the three P. ledebouriana populations and one population of P. tenella were measured for plant height. The results clearly showed that samples from the two species have a distinct separation based on this trait (P < 0.0001). Similar results for different morphological characteristics confirm the difference among other representatives of the genus (Devi, Singh & Thakur, 2018). Hence, evaluating plant height can be a reliable way of distinguishing between two closely genetically related species within the section Amygdalopsis. The performance of this trait in plants is reliably related to the elevation above sea level (a.s.l.), as the lowest altitude for P. ledebouriana samples was higher than 1.7 m a.s.l. and the highest altitude for P. tenella samples was lower than 1.5 m a.s.l. (Table S1). We assessed samples from two species using 19 polymorphic SSR loci to confirm the conclusion based on the plant height study. The application of SSR markers resulted in a constructed UPGMA phylogenetic tree that separated 20 samples of P. tenella (4-UK) from 60 samples of P. ledebouriana (1-UR, 2-KO, 3-KA) (Fig. 4). This result was also supported by the PCoA plot, where PC2 (41.2%) split 4-UK samples from 1-UR and 2-KO.

In contrast, PC1 (49.1%) separated 3-KA from the remaining three populations (Fig. 3). Interestingly, the clusterization in Fig. 4 suggests that, generally, there is a low level of admixture between populations, which supports the model of “isolation by distance” (Meirmans, 2012). The genetic heterozygosity index (Nei’s index) assessment showed that the highest genetic diversity was registered in the population of P. tenella (0.606). In contrast, the lowest index was recorded in population 3-KA (0.449), representing the area with the highest sampling elevation. High altitude is a sufficiently solid environmental factor that negatively influences genetic variation in P. ledebouriana. Nevertheless, the separation of 3-KA from 1-UR and 2-KO supported a more significant genetic variation within the species. The high genetic differentiation between mountain and steppe populations is most likely due to the factor of the steppe zone and the presence of anthropogenic pressure on the P. tenella (4-UK) population. In turn, mountain populations are distinguished by mountain isolation of ridges. The analysis of samples by DNA genotyping using SSR markers suggested that five loci were characterized as markers with the highest polymorphism level (Table 3). These five SSR loci could also be recommended for the discrimination of Prunus species in other studies. In addition, a t-test was applied to test the association of 19 polymorphic SSR loci with plant height (Table 4). It was concluded that nine out of 19 SSR loci were significantly associated with plant height. This result may not be a direct reflection of associations between SSRs and plant height but an indication of the genetic differences between P. ledebouriana and P. tenella, as these two species have significantly differed in plant height (Fig. 1). Therefore, these nine SSR loci can be efficiently used in further studies of discrimination between P. ledebouriana and P. tenella. The assessment of the population structure using the STRUCTURE package suggested that populations of two species started separating at steps K = 3 and K = 4, which is another indication that P. ledebouriana and P. tenella are two different species. The evaluation of samples in four clusters at K = 4 (Fig. 3) showed little admixture level, supporting the model of isolation by distance with a limited gene flow among the populations. Mantel tests revealed a positive correlation between geographic and genetic distance among populations (r = 0.4387), demonstrating consistency with the isolation-by-distance model.