Persian walnut (
Seed-propagated walnut trees have distinct morphological and phenological characteristics, in particular growth and bearing habit, tree vigour, bud break, kernel ratio and colour and nut weight, which are all highly variable (Sutyemez, 2006; Şimsek, 2010; Ertürk et al., 2014; Kirca et al., 2014). Commercial plantations of walnuts are relatively new to Turkey and are the result of programmes implemented by the Turkish government to encourage increasing walnut production. Successful commercial orchards were established in the last two decades mostly in the Aegean, Marmara, Southeastern Anatolia and Mediterranean regions, with both domestic and foreign cultivars. ‘Chandler’ is becoming the most popular cultivar, due to consumer preferences. However, it alone cannot meet the demands for high-quality walnuts in Turkey (Keles et al., 2014). The introduced cultivars such as Chandler did not adapt well to different agricultural regions in Turkey (Keles et al., 2014). On the other hand, in Turkey, non-grafted walnut seedlings had also been used in new walnut plantations. This has provided the opportunity to select the best genotypes from natural populations for good yield and nut quality characteristics. Yalova-1, Yalova-2, Yalova-3 and Yalova-4 as well as Sebin and Bilecik were the first Turkish walnut cultivars to be selected from natural walnut populations in different regions in Turkey (Olez, 1976). The Maras-18 cultivar is selected from Kahramanmaraş province in Turkey with a higher yield and better quality nuts than the mentioned cultivars and it presents homogamy in flowering (Sutyemez, 1998). Kaman-1 is another cultivar selected from natural population in inner Anatolia with a high yield and good nut characters. The Maras-18, Şebin and Bilecik cultivars are currently preferred in areas with short vegetation periods due to their early leafing time (Keles et al., 2014).
The main contributions of seed-propagated genotypes to plant breeding have been traits for more efficient nutrient uptake and utilisation, as well as useful genes for adaptation to stressful environments such as water stress, salinity and high temperatures. Seed-propagated walnut genotypes are generally adapted to local environment better than introduced cultivars (Keles et al., 2014). With the continued cycles of local planting, harvesting and farmers’ selection, over time, genotypes will be selected for local environmental and agro-ecosystem conditions and practices, just as ecotypes of wild species are adapted to the local environmental conditions. Seed-propagated genotypes are adapted to their growing conditions and they possess adaptive complexes associated with the special conditions of cultivation, pure-stand associations, harvesting and other factors (Bennett, 1970; Villa et al., 2006).
The walnut tree is characterised by differentiation, adoption and a long history of cultivation throughout the temperate regions of the world. Due to these characteristics, it has formed a wide genetic diversity under appropriate environmental conditions (Martinez et al., 2010). However, there is only limited research on the identification of this extensive diversity and differentiation in seed-propagated walnut germplasm.
China is the world leader in walnut production (accounting for 50% of the total) with 1,060,000 metric tons of walnut production annually, followed by the USA. The European Union, Ukraine, Chile, Turkey and Moldova are the other top regions/countries in walnut production (FAO, 2018). According to the agricultural statistics, there are 17,000,000 trees on 93,000 ha in Turkey corresponding to 210,000 tons of walnut production.
Identifying walnut genotypes and deter mining genetic relationships are necessary for registering genotypes as cultivars (Ercisli et al., 2012; Keles et al., 2014; Kirca et al., 2014). The traditional identification method by examining morphological features requires skilled professionals, which corresponds to the cost of genetic analysis. Furthermore, the approach is time-consuming and laborious, and it depends on the season (Ercisli et al., 2008; Guney et al., 2018). Conversely, DNA molecular marker technology can be used without the influence of environmental conditions and is relatively quick, inexpensive and extremely accurate in identifying varieties (Eyduran et al., 2016; Poczai et al., 2013). The technology has been widely applied in the analysis of genetic diversity, variety identification and establishment of fingerprints in various fruit trees (Ercisli et al., 2008; 2011; Sutyemez, 2006; Guney et al., 2019). Various types of molecular markers have been widely applied in walnut genetic diversity studies. Random amplification of polymorphic DNA (RAPD) (Dogan et al., 2014), simple sequence repeats (SSRs) (Foroni et al., 2007; Karimi et al., 2010; Chen et al., 2014; Bernard et al., 2018; Khokhlov et al., 2018), inter-simple sequence repeat amplification (ISSR) (Potter et al., 2002), amplified fragment length polymorphism (Kafkas et al., 2005; Bayazit et al., 2007) and single nucleotide polymorphism (SNP) analysis (Ciarmiello et al., 2011) were utilised in examining genetic variability and identity of walnut germplasm resources.
Recently, genome and transcriptome analysis had been done on
In Turkey, particularly in eastern Anatolia region, walnut production is still carried out using seed-propagated genotypes. Thus, walnut fruits in the market are mixture of fruits from various trees. Therefore, walnut industry and consumers are challenged in getting products of standard quality. Further breeding investigations on different seed-propagated walnut trees are required for selection of genotypes with high yield and quality. The success of breeding programmes depends mainly on the available genetic information of the breeding material. In addition, fast and accurate identification of genotypes and evaluation of the genetic variability provide substantial benefits for the certification of cultivars with specific characteristics. In this study, SSR markers were used to analyse the genetic relationships of seed-propagated walnut genotypes and to reveal their genetic diversity at molecular level.
To our knowledge, this is the first study on SSR analysis of seed-propagated walnuts grown in Eastern Anatolia region of Turkey. Therefore, the objectives of this research were to characterise and evaluate the genetic diversity of 32 seed-propagated walnut genotypes and 2 standard cultivars using SSR markers and to develop strategies for preserving endangered genetic resources of this species.
In total, 32 seed-propagated walnut genotypes along with 2 standard cultivars (‘Sebin’ and ‘Bilecik’) from Kagizman district of Eastern Anatolia region in Turkey (Table 1) were analysed using SSR markers. Two standard cultivars (‘Sebin’ and ‘Bilecik’) were previously selected among seed-propagated walnut genotypes. Kagizman district has a narrow geographic range of longitude (40°14N–43°11E) located in Aras valley of eastern Turkey. In this valley, walnut populations consist of old walnut trees from open pollinated seedlings (70-to 120-year-old trees). First, the mature seedling trees origin was labelled based on their region area (KW1 to KW32). The selected genotypes were healthy and had a full crop. Growth habit, bearing habit, tree vigour, bud break, nut weight, kernel ratio and kernel colour were determined (Table 1). Each genotype (KW1 to KW32) was presented by one seed-propagated tree. The trees cultivated in the valley represent the local population (seedlings) that were randomly planted by humans or birds. The ones with better yield and fruit characteristics among these walnut trees were selected and kept by the people, and they were considered as endangered genetic diversity because they were propagated by seeds. Growth habit, bearing habit, tree vigour and bud break were determined according to IPGRI (1994) in 2016. During harvest season, 50 nuts from each tree were randomly collected and evaluated for nut characteristics (weight, kernel percentage and kernel colour). The weights of nut and kernel were measured using an electronic balance with 0.01 g precision. Kernel percentage was estimated using formulas ‘kernel weight/nut weight × 100’ (Eriksson, 1998).
Some important morphological traits of walnut genotypes
Genotypes | Growth habit | Bearing habit | Tree vigour | Bud break | Nut weight (g) | Kernel ratio (g) | Kernel colour |
---|---|---|---|---|---|---|---|
KW1 | Semi upright | Lateral | Medium | Late | 9.28 | 51.4 | Light amber |
KW2 | Spreading | Intermediate | Medium | Late | 10.25 | 47.6 | Light |
KW3 | Upright | Lateral | Strong | Mid Early | 10.86 | 56.4 | Light |
KW4 | Semi upright | Lateral | Medium | Early | 8.76 | 55.3 | Light |
KW5 | Spreading | Intermediate | Strong | Late | 12.43 | 48.7 | Light amber |
KW6 | Semi upright | Lateral | Medium | Late | 11.78 | 62.5 | Light |
KW7 | Semi upright | Lateral | Medium | Early | 9.05 | 59.3 | Light |
KW8 | Spreading | Lateral | Medium | Mid Early | 13.20 | 48.3 | Light amber |
KW9 | Upright | Lateral | Weak | Late | 12.56 | 45.3 | Light amber |
KW10 | Semi upright | Lateral | Medium | Late | 8.94 | 59.9 | Light |
KW11 | Spreading | Lateral | Strong | Early | 11.46 | 55.9 | Light |
KW12 | Semi upright | Intermediate | Medium | Late | 9.44 | 56.9 | Light amber |
KW13 | Upright | Lateral | Weak | Early | 11.28 | 53.4 | Light |
KW14 | Semi upright | Lateral | Strong | Late | 11.67 | 60.4 | Light amber |
KW15 | Spreading | Lateral | Strong | Mid Early | 9.60 | 62.3 | Light |
KW16 | Semi upright | Lateral | Medium | Early | 12.54 | 43.8 | Light |
KW17 | Upright | Lateral | Medium | Late | 10.89 | 47.3 | Very light |
KW18 | Semi upright | Intermediate | Medium | Mid Early | 9.76 | 60.1 | Light amber |
KW19 | Spreading | Lateral | Medium | Late | 9.28 | 51.4 | Light amber |
KW20 | Spreading | Lateral | Strong | Early | 10.20 | 42.8 | Very light |
KW21 | Spreading | Lateral | Medium | Late | 11.55 | 61.8 | Light |
KW22 | Semi upright | Lateral | Medium | Late | 9.09 | 50.6 | Light amber |
KW23 | Semi upright | Intermediate | Weak | Early | 9.33 | 51.0 | Light |
KW24 | Semi upright | Lateral | Medium | Late | 12.44 | 56.3 | Light |
KW25 | Upright | Intermediate | Strong | Early | 9.80 | 55.0 | Light amber |
KW26 | Semi upright | Lateral | Medium | Late | 10.50 | 44.4 | Light amber |
KW27 | Semi upright | Lateral | Strong | Mid Early | 9.15 | 52.7 | Light |
KW28 | Spreading | Lateral | Strong | Early | 11.75 | 50.0 | Light |
KW29 | Semi upright | Lateral | Medium | Early | 10.80 | 57.0 | Light amber |
KW30 | Semi upright | Lateral | Medium | Late | 12.65 | 48.1 | Light |
KW31 | Semi upright | Lateral | Medium | Late | 9.28 | 51.4 | Light amber |
KW32 | Upright | Intermediate | Weak | Early | 8.88 | 52.5 | Light |
Sebin | Semi upright | Lateral | Medium | Late | 11.80 | 58.9 | Light amber |
Bilecik | Semi upright | Lateral | Medium | Late | 9.19 | 51.4 | Light amber |
Genomic DNA was extracted from young leaf tissues using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA) according to the instr uctions provided by the manufacturer. Subsequently, an RNAse treatment was performed to the eluted DNA samples. Purity and concentration of the DNA were confirmed both on 1% (w/v) agarose gels and by NanoDrop ND-1000 Spectrophotometer.
A total of 21 SSR loci were used (Table 2). Polymerase chain reaction (PCR) was conducted in a volume of 10 mL, containing 15 ng genomic DNA, 5 pmol of each primer, 0.5 mM dNTP, 0.5 unit GoTaq DNA polymerase (Promega) and 1.5 mM MgCl2. The forward primers were ‘labelled’ with WellRED fluorescent dyes D2 (black), D3 (green) and D4 (blue) (Proligo, Paris, France). Reactions without DNA were included as negative controls. PCR amplification was performed using the Biometra PCR System. The amplification conditions consisted of an initial denaturation step of 3 min at 94°C, followed by 35 cycles of 1 min at 94°C, 1 min at 52°C–56°C and 2 min at 72°C with a final extension at 72°C for 10 min. The PCR products were run on CEQTM 8800 XL Capillary Genetic Analysis System (Beckman Coulter, Fullerton, CA) to determine polymorphisms. The analyses were repeated at least twice to ensure reproducibility of the results.
SSRs, number of detected alleles, observed heterozygosity (
SSR primers | References | Number of alleles ( | Observed heterozygosity ( | Expected heterozygosity ( | PIC |
---|---|---|---|---|---|
WGA001 | Dangl et al. (2005) | 7 | 0.47 | 0.58 | 0.67 |
WGA004 | Woeste et al. (2002) | 5 | 0.64 | 0.60 | 0.55 |
WGA005 | Woeste et al. (2002) | 6 | 0.80 | 0.71 | 0.77 |
WGA009 | Dangl et al. (2005) | 6 | 0.47 | 0.65 | 0.68 |
WGA027 | Woeste et al. (2002) | 3 | 0.55 | 0.49 | 0.72 |
WGA032 | Woeste et al. (2002) | 8 | 0.80 | 0.67 | 0.75 |
WGA054 | Woeste et al. (2002) | 9 | 0.53 | 0.60 | 0.58 |
WGA069 | Woeste et al. (2002) | 7 | 0.71 | 0.60 | 0.64 |
WGA071 | Woeste et al. (2002) | 4 | 0.52 | 0.49 | 0.69 |
WGA072 | Woeste et al. (2002) | 5 | 0.74 | 0.64 | 0.60 |
WGA079 | Woeste et al. (2002) | 4 | 0.58 | 0.53 | 0.55 |
WGA089 | Dangl et al. (2005) | 5 | 0.62 | 0.71 | 0.79 |
WGA118 | Dangl et al. (2005) | 5 | 0.58 | 0.80 | 0.65 |
WGA202 | Dangl et al. (2005) | 8 | 0.70 | 0.63 | 0.83 |
WGA225 | Dangl et al. (2005) | 5 | 0.41 | 0.64 | 0.62 |
WGA276 | Dangl et al. (2005) | 11 | 0.39 | 0.58 | 0.85 |
WGA321 | Dangl et al. (2005) | 8 | 0.70 | 0.60 | 0.54 |
WGA331 | Dangl et al. (2005) | 5 | 0.58 | 0.41 | 0.61 |
WGA332 | Dangl et al. (2005) | 6 | 0.68 | 0.61 | 0.58 |
WGA349 | Dangl et al. (2005) | 8 | 0.55 | 0.74 | 0.76 |
WGA376 | Dangl et al. (2005) | 10 | 0.52 | 0.59 | 0.80 |
Average | 6.43 | 0.60 | 0.62 | 0.68 |
PIC, polymorphic information content; SSR, simple sequence repeats.
The genetic analysis software ‘IDENTITY’ 1.0 (Wagner and Sefc, 1999) was used according to Paetkau et al. (1995) for the calculation of number of alleles (
The genotypes and standard cultivars used in this study were not previously subjected to molecular characterisation. The co-dominant inheritance and the distribution throughout the genome provide easy manipulation and large allelic variation making microsatellite analysis the optimal method for assessing genetic diversity (Suprun et al., 2013). The genetic diversity analysis of walnut landraces and standard cultivars included the average number of alleles per locus (
In total, 135 alleles were obtained from the 21 SSR loci analysed. Results show that the number of alleles per locus ranged from 3 (WGA027) to 11 (WGA276) with a mean number of 6.43 alleles per locus (Table 2). Along with WGA276 locus, WGA376 (with 10 alleles per locus) and WGA054 (with 9 alleles per locus) yielded the highest number of alleles (Table 2). WGA027, WGA071 and WGA079 gave the lowest alleles per locus (Table 2). Walnut cultivars and landraces originating from different parts of the world were previously characterised by SSR analysis. Ruiz-Garcia et al. (2011) used 57 common walnut cultivars, mainly from Spain and the USA, in SSR analysis. They obtained a total of 97 alleles and an average of 5 alleles per locus. Their study obtained the highest number of alleles per locus from WGA032 (10 alleles per locus) and WGA227 (9 alleles per locus), which is in accordance with our results. Khokhlov et al. (2018) used 15 walnut cultivars from Ukraine for SSR analysis and found that the number of alleles per locus ranged from 5 (WGA69) to 13 (WGA276), with an average of 8 alleles per locus indicating a high degree of polymorphism and a high genetic diversity within samples. In the characterisation of 44 walnut genotypes by 14 microsatellite loci reported by Dangl et al. (2005), the number of alleles per locus ranged from 3 to 8 with an average of 5.2. Karimi et al. (2010) studied diversity of natural populations of
The expected and obser ved heterozygosities were found to be between 0.49 (WGA027) and 0.80 (WGA118), and 0.41 (WGA225) and 0.80 (WGA005), respectively (Table 2). The average expected and observed heterozygosities were found as 0.62 and 0.60, respectively. Results clearly indicate that our seed-propagated outcrossing walnut genotypes have shown a high degree of heterozygosity. Victory et al. (2006) reported a high observed heterozygosity value (higher than 0.80) of wild
PIC refer to the value of a marker for detecting polymorphism within a population, depending on the number of detectable alleles and the distribution of their frequency; thus, it provides an estimate of the discriminating power of the marker and therefore the PIC values were calculated for all primers (Table 2). The average PIC value was obtained as 0.68. The most informative locus was WGA276 (0.85), followed by WGA202 (0.83), WGA376 (0.80) and WGA089 (0.79), whereas the least informative loci were WGA321 (0.54), WGA004 (0.55) and WGA079 (0.55). According to our results, 13 loci are classified as informative markers (PIC > 0.5) and 8 loci as suitable for mapping (PIC > 0.7) (Table 2). These results indicate that all the markers could provide valuable information on walnut genetics and breeding research. The results also show that the walnut populations of Kagizman district have relatively high levels of genetic diversity. Pollegioni et al. (2011) observed that except WGA004 (0.355) and WGA331 (0.382), all markers had PIC > 0.50. Vahdati et al. (2015) analysed 6 walnut populations from Iran using 17 microsatellite loci and found that PIC for the loci ranged from 0.56 to 0.82 with an average of 0.72. The PIC measure is an important component in the planning of breeding programmes, and it is a key information element and statistical indicator (Chesnokov and Artemyeva, 2015).
The genetic relationship between the genotypes was clearly depicted in the dendrogram that was constructed from the DNA profile. The dendrogram was obtained after processing the experimental data. The dendrogram divided the analysed sample into two large clusters, indicating the heterogeneity of the seed-propagated walnut genotypes in the Kagizman district in Turkey. The node separating cluster I from cluster II had a high bootstrap value indicating high variation between the genotypes in clusters I and II (Figure 1).
The UPGMA dendrogram based on simple matching similarity matrix obtained using 21 SSR markers, illustrating the relative similarity among 32 walnut genotypes and 2 standard cultivars.
The first cluster consisted of 10 seed-propagated genotypes and 2 standard cultivars. Within the first cluster, two subgroups were formed. The first subgroup included 10 seed-propagated genotypes. The second subgroup consisted of two seed-propagated genotypes and two standard cultivars. The closest genotypes in the first cluster were KW18 and KW31 with 0.83 similarity ratio. KW18 and KW31 have some morphological similarities such as semi upright growth habit, medium tree vigour, similar nut weight and light amber kernel colour (Figure 1). It is worth noting that KW18 and KW31 showed less diversity suggesting homogeneity and genetic flux or common origin. Within cluster I, cvs. ‘Bilecik’ and ‘Sebin’ were grouped together with 0.55 similarity ratio. ‘Bilecik’ and ‘Sebin’ were closely clustered with the seed-propagated genotypes (KW10 and KW24). This could be explained by the presence of a common genetic origin among the genotypes/cultivars despite their great diversity. The second cluster consisted of 20 seed-propagated genotypes and further divided into 2 subgroups. Within cluster II, the first subgroup included 12 genotypes and the second subgroup contained 8 genotypes. Genotypes KW27 and KW28 were the closest with 0.87 similarity ratio, while KW29 and KW22 were the most distant with 0.23 similarity ratio (Figure 1). KW27 and KW28 show some similar morphological characteristics. Both genotypes have lateral bearing habit, strong tree vigour, similar kernel ratio and light kernel colour. The SSR similarity level found between KW27 and KW28 indicates a common or related ancestry. In addition, the genotypes that have upright growth habit (KW2, KW13, KW17, KW25 and KW32) except KW9 were assigned in cluster II. Certain degree of relationship between walnut genotypes is probably related to their seed-propagated nature.
According to these results, we could infer that the 34 walnut genotypes/cultivars are divergent. The characteristics of seed-propagated walnuts in relation to the magnitude of allelic and genetic diversity in contrast to cultivars are considered to be significantly more genetically diverse. Thus, a seed-propagated walnut genotype is highly variable in appearance and highly genetically diverse or genetically heterogeneous. They are recognised as a distinct entity via common-shared traits. These traits will allow the distinction of one seed-propagated genotype from another or from modern cultivars for the same crop (Sutyemez, 1998; Şimsek, 2010).
There were some similarities between morphological traits and molecular data. The genotypes assigned in cluster I have semi upright growing habits, lateral bearing, medium tree vigour and late bud-breaking characteristics (Figure 1). However, the genotypes in cluster II also have the above-mentioned morphological traits. All genotypes and two cultivars examined in this study are seed-propagated and with unknown parentage.
Such a wide interval of similarity values indicates a wide range of genetic diversity of the analysed walnut genotypes. The seedling-originated genotypes also affect the present results. About 5 million walnut trees in Turkey are composed generally of non-grafted trees, thus increasing the genetic variation of walnut germplasm and offering an opportunity for breeders to select the best genotypes. Genotypes collected from various parts of Kagizman district showed diverse clustering which could be due to the outcrossing nature of walnut. Ahmed et al. (2012) reported similarity between 0.12 and 0.78 with average of 0.49 in walnut through SSR analysis. Dogan et al. (2014) reported genetic similarity values from 0.58 to 0.91 in walnut cultivars from Turkey.
PCA was applied to the raw data from the SSR ‘1’ and ‘0’ matrix by using NTSYS 2.10e software, and the contribution rates of the first 3 principal coordinates were 34.44%, 27.15% and 12.80%, respectively, accounting for 74.39% of the genetic similarity variance. These walnut genotypes/cultivars samples were partitioned into two distinct groups. PCA Group 1 included 12 samples, and PCA Group 2 included 19 samples (Figure 2) which was in general consistent with the cluster analysis output. PCA1 appeared to be much more informative, clearly separating two different groups. The genetic variation was higher in Group 2 than in Group 1. The markers employed in this study will be useful for the characterisation and comparison of walnut germplasm collections and for the detection of propagation errors. The evaluation of the molecular diversity of walnut genetic resources is important for optimal programme development aiming conservation. These markers were able to identify uniquely all the tested walnut genotypes.
Two-dimensional PCA of 34 individual walnut genotypes/cultivars.
Detailed characterisation of plant genetic structure is a vital step of a breeding programme, together with efficient protection and use of plant genetic resources. Molecular marker analysis is considered as the most reliable and powerful method for such characterisation. To date, only a limited number of studies have examined the genetic relationships between seed-propagated walnuts using SSR-based approach. Understanding genetic variation and relationships within genotypes and standard cultivars is crucial for fundamental research, conservation and the potential utilisation of genetic resources for walnut breeding. The results of this study demonstrated that genetic diversity exists between the investigated walnut genotypes. SSRs revealed a high mean number of alleles per locus, as well as a high heterozygosity and PIC values. The study outputs indicated that SSR markers can be successfully used to determine the diversity of seed-propagated walnut genotypes. Genetic variations among seed-propagated walnut genotypes could be useful in selecting parents to be crossed to generate suitable populations intended for breeding strategies. The information provided by this research will shed light on future studies of walnut cultivar selection, planting and production.