Morphologic and molecular assessments of cucumber (Cucumis sativus L.) landraces

In this study, 90 locally grown cucumber (Cucumis sativus L.) landraces were collected and morphologically characterized using 20 descriptors derived from UPOV (International Union for the Protection of New Varieties of Plants). Genetic diversity and relationships of the genotypes were revealed using 20 sequence-related amplified polymorphism (SRAP) marker combinations. The discrimination power of each polymorphic marker (estimated by the polymorphism information content) ranged from 0.15 to 0.99 with an average of 0.73. Dice's similarity coefficient ranged between 0.00-1.00. The cluster analysis that was conducted using the unweighted pair group method of arithmetic averages (UPGMA) for both molecular and morphologic data showed that all of the genotypes fell into two main groups and many subdivisions. According to morphological data, fruit length, diameter and weight of the genotypes were determined between 6.5 32.5 cm, 25 52 mm and, 28 625 g respectively. It is clear from the results, a moderate level of genetic diversity, which has the potential for broadening the genetic base, was observed among the Turkish cucumber landraces.


Introduction
Cucumber (Cucumis sativus L.) is grown widely in open field or glasshouse conditions. It has modest levels of vitamin A, C, and, some phenolic compounds known to have antioxidant properties (Kumar et al., 2010). India is the primary center of diversity for cucumber (Wehner and Robinson, 1991). It has a narrow genetic base with a genetic variability of 3-8% so it is more susceptible to biotic and abiotic stresses (Rao and Hodgkin, 2002).
Nowadays, climate change threatens to reduce crop yield and harming the food security. Local landraces have adaptation skills to shifting climatic conditions. They have been a major resource for crop breeding, due to their tolerance to the extreme climate changes and the resistance to some pests and diseases (Rao and Hodgkin, 2002). Their genetic and phenotypic features need to be characterized carefully for the proper use in breeding programs (Parisi et al., 2017). The success of any plant breeding program depends on the genetic variability (Bartaula et al., 2019) Establishing a gene pool and determination of their relationships is the first step for plant improvement (Normohammadi et al., 2017).
Morphological markers are good determinants but they are not stable and they influenced by ecologic conditions (Fares et al., 2009). Molecular markers are not affected by the environmental changes and they serve as a tool to overcome the deficiencies of morphological markers in the characterization studies (Agarwal et al., 2008;Gostimsky et al., 2005;Gulsen and Mutlu, 2005). For this reason, using morphological and molecular marker systems together is preferred for the better clarification of the relationships between the genotypes.
In this research, 20 descriptors which were derived from "International Union for Protection of New Plant Varieties" (UPOV) were used for the morphological determinations and the most polymorphic 20 SRAP marker combinations were used for the molecular characterization. SRAP is a PCR-based technique. It depends on the amplification of open reading frames (ORFs) and two-primer amplifications. The primer sequences are random and they have 17-21 nucleotides in length (Agarwal et al., 2008).
Turkey is located on the transition zones. Its geographic position and the variety of agro-climatic conditions caused great biodiversity. This research aimed to establish a characterized gene pool formed by cucumber landraces cultivated in different regions of Turkey. Even though many studies on Turkish cucumber diversity have been completed earlier (Gozen, 2008;Aydemir, 2009), for the first time with this research Turkish landraces were collected and characterized using morphologic and molecular markers together. Therefore, obtained results are significant both for the local and international researchers.

Material
This study was conducted at Bati Akdeniz Agricultural Research Institute-Antalya located between 36°55'0.98" N 31°05'56.00" E coordinates. Ninety landraces were collected from 52 different areas, used as plant material. The information about their given codes, cultivation areas, and geographic locations were presented in Table 1. Thirty-six accessions came from the Genebank of Aegean Agricultural Research Institute and the other 54 accessions collected from cucumber cultivation areas via agricultural district offices before starting to this study.

Methods
Gene pool establishment Collected genotypes were selfed for two generations to avoid risks because of possible seed mixture. For this reason, all seeds were planted in 2017 spring and autumn season. When the seedlings at the 2-3 leaf stage, they were transplanted to the greenhouse. During the vegetation period, watering, fertilizing, and diseases and pest management were carefully followed. At the fruiting stage, mixed accessions were detected; they were considered as another genotype and coded differentially. Pollination studies were made in each population separately. Hand pollinated flowers were tagged with coloured materials. Following pollination, after 50-55 days, ripen fruits were harvested; seeds were washed for cleaning and dried then kept in cooled storage until the next season.

Morphological characterization
Spring and autumn period of 2018, morphological characterization of the genotypes was carried out according to the descriptors were given in Table 2. Five plants per accession were characterized, with 20 descriptors, eight of them referred to plant traits and 12 related to the fruit.

DNA extraction
Selfed seeds of 90 genotypes were sown again in the spring of 2019. When the seedlings were at 2 -4 leaf stage, plant samples were taken to perform DNA extraction. The DNAs of plants were isolated according to the modified CTAB protocol by Doyle and Doyle (1990). After the extraction, the DNA concentrations were determined in 1% agarose gel. All extracted genomic DNAs were stored at -20 o C in TE buffer for SRAP analysis.  Gozen (2008) Features

SRAP analysis
The most polymorphic 20 SRAP primer combinations were determined first before starting this stage. The primers were obtained from the previous studies on cucumber conducted by Ferriol et al. (2003;. SRAP primers and their sequences that were used to create 20 combinations were presented in Table 3. The PCR reaction solution was composed of 1 µl DNA (20 ng DNA), 15 µl 1 x PCR master mix, 1 µl each primer (0.3 µM each primer) and 7 µl ddH2O. PCR amplification was 5 minutes for initial denaturation at 94 °C, then 5 cycles composed of denaturing at 94 °C for 1 minute, annealing at 35 °C for 1 minute, extension at 72 °C for 1 minute and followed by 35 cycles; heating at 94 °C for 1 minute, annealing at 55 °C for 1 minute, extension at 72 °C for 1 minute and 10 minutes for the final extension at 72 °C. PCR amplifications were repeated twice for each primer combination.
The PCR products were fractionated on 2.5% high-resolution agarose gel in 1XTAE buffer at 110 V for 3 hours and photographed (Kodak Gel Logic 200) under UV light for further analysis. A 100 bp DNA ladder was used as a molecular standard.

Data analysis
Genetic similarity and phylogenetic relationships were analysed by the UPGMA (Un-weighted pairgroup method, arithmetic average) clustering procedure using the software NTSYS (Numerical Taxonomy Multivariate Analysis System) version pc 2.2 (Rohlf, 1998). For morphological data, each genotype was characterized with descriptors were shown in Table 2. For molecular analysis data, each genotype was identified for each primer based on the presence (1) and absence (0) of bands. Similarity indices were calculated according to Dice (1945). Polymorphism rates (Pr) = (Polymorphic allel / Total allel) X 100 and polymorphism information content (PIC) = 1-Σ fi2 (Smith et al., 1997) were calculated.

Results and Discussion
In this study collected local cucumber genotypes were determined morphologically different. Morphological observations of 47 genotypes using 10 UPOV descriptors were presented in Table 4 as an example. According to this data, genotypes were classified in terms of their growth habit, 28 of them (31%) determined as strong, 49 of them (54%) were medium and 13 of them (14%) were reported as weak. Poor growth leads to plants susceptible to diseases, pests, and harsh environmental conditions. Therefore, while a breeding program set up, plant growth habits should be considered. Seventy-eight of used genotypes were detected as monoecious, regarding the flowering time, lateness was observed on 54 genotypes and 19 of them have no prickliness.
Although cultivated cucumbers are mostly similar in terms of root and stem development, when the fruit characteristics, especially fruit shape is considered they could be quite diverse (Bisognin, 2002). In the current study, fruit length, diameter, and weight of the genotypes were determined between 6.5 -32.5 cm, 25-52 mm, and 28-625 g respectively. Features such as fruit length, diameter, and color are used to determine the economic importance of a variety (Kennard and Havey, 1995;Ahmed et al., 2004).
It was understood from the study, while the cucumber is native to India where its wild forms still exist (Sebastian et al., 2010), examined landraces are morphologically quite diverse because of Turkey's location between the transition zones ( Figure 1). Genotypes were also observed regarding their number of lateral shoots in the first 15 nodes. While 11 genotypes have no lateral shoots, 79 genotypes have lateral shoots between 4 and 14. In a study, Wehner and Guner (2002) reported that, some genotypes achieve high yielding with branching ability while others reach it with more nodes. Also, the higher yield of some genotypes in their study was associated with both extra branches and nodes.  In this study, 20 morphologic descriptors which were derived from UPOV by Gozen (2008) used for the definition of the phylogenetic relationships among the 90 Turkish cucumber landraces. Obtained morphological data was used to draw the dendrogram (Figure 2) generated by the UPGMA method. According to the dendrogram presented in Figure 2, two major and four subgroups were revealed with the similarity rate of %25. Although the most distant genotypes were determined YG-2 and YG-34 gathered from Canakkale, the most similar ones were YG-61 from Mersin, YG-67 from Antalya ( Figure 3). Thus, the fact that, the origins of genotypes were close or far apart did not affect their morphological similarity in the research. Determination of the genetic diversity of a gene pool is crucial for the success of the breeding programs. Because they are more reliable and consistent, using neutral markers is essential in such studies (Al-Rawahi et al., 2011). Therefore, morphological and molecular markers were preferred in the current study and the discrimination power of each polymorphic marker (estimated by the polymorphism information content) ranged from 0.15 to 0.99 with an average of 0.73. Dice's similarity coefficient ranged from 0.00 to 1.00. The dendrogram provided from the molecular data was given in Figure 4. Accordingly, 90 genotypes used in the study divided into 2 main clusters which were subdivided into themselves. Genetic diversity and relationships of the genotypes were revealed using 20 sequence-related amplified polymorphism (SRAP) marker combinations.
According to the dendrogram presented in Figure 4, there were 37 genotypes in cluster B, remaining 45 genotypes were found in cluster A. According to Dice similarity and distance indexes group A consisted of genetically the most similar genotypes with the smallest similarity value of 0.88. Mliki et al. (2003) studied the genetic diversity of 26 African cucumber accessions. Their data suggest that some genotypes from Egypt have unique genetic diversity and this germplasm has the potential for broadening the genetic base of cucumber breeding studies. Gaikwad (2011) reported that genotypic variation was slightly low as compared to phenotypic variation in cucumber. The highest genotypic, as well as phenotypic variation was observed on fruit length, fruit number, and fruit weight. With the current study, it was cleared that Turkey has reasonable genetic diversity in terms of cucumber and the 20 SRAP primer combinations established here could be used for germplasm identification in further classification studies. Today management of biodiversity for sustainable food has become a significant subject. Landraces of every country represent a major source of diversity because of the adaptation capability to the habitats of origin. These resources could be used in the plant breeding programs for the constitution of more resilient and qualitatively improved varieties (Parisi et al., 2017).
In Table 5, used 20 SRAP primer combinations and their number of total alleles, polymorphic alleles, polymorphism rates (%), and polymorphism information content (PIC) values were presented. Considering these, the most allelic primer combination was found as me11-em9 with the total 12 alleles. The least allelic primer combinations were determinate as me12-em10 and me6-em16 with the 5 alleles. Polymorphism rate (%) was ranked between %16-%100.

Conclusions
With this project, 90 of the local cucumber genotypes grown in different regions of Turkey were collected and morphologically characterized considering the 20 UPOV descriptors. Later, genetic similarity of the genotypes was identified between the 0.00 -1.00 using 20 SRAP primer combinations showing polymorphism. Used marker systems also allowed the efficient management of the gene pool by eliminating redundant accessions. It was cleared that Turkey has reasonable diversity regarding cucumber. It was also understood that the obtained gene pool has a potential for broadening the genetic base of cucumber for further breeding studies. According to the clusters achieved in the study, morphologically similar and molecularly distinct genotypes could be used as parent lines in the further cucumber breeding studies.