Application of Microsatellite and Inter Simple Sequence Markers for Certifying Trueness to Type of Grafted Ramets and Clustering of Persian Walnut Varieties

The utility of seventeen Microsatellite (SSR) markers and fteen inter simple sequence repeats (ISSR) markers for the identication of twenty eight ramets of 11 varieties of walnut (Juglans regia) was explored. Thirty nine individual genomes were screened using 61 and 38 scorable fragments from SSR and ISSR markers, respectively. The least polymorphic SSR locus was WGA004 (two alleles) and the most polymorphic (5 alleles) was WGA276. Polymorphism information content values ranged from 0.08 (WGA004) to 0.43 (WGA032) in SSR markers and from 0.11 (AGA (AC)7) to 0.49 (CAC(TGT)5) in ISSR markers, with an average of 0.29 and 0.19, respectively. In most cases, grafted varieties with identical names also had the same microsatellites prole. The principal coordinate analysis and clustering (UPGMA) based on the combined marker set emphasized two failures in grafting or off-types, ramets identied as Serr 4 (S4) and Vina 1 (V1). The presence of two off-type ramets in the walnut research orchard emphasizes the importance of using molecular certication for proving true-to-type of walnut orchards. Using 13 polymorphic SSRs, we tabulated a DNA ngerprint chart of 11 walnut varieties. Except for ‘Chandler’, each cultivar could be distinguished using a combination of only two SSR loci. The 13 SSRs markers evaluated in this study could be used in future to identify clones produced from the varieties.


Introduction
Persian walnut (Juglans regia) is widely cultivated for nut and wood production (McGranahan and Leslie 2012). Originally cultivated in Persia's plateau, it has been introduced all around the world (McGranahan 2007;Pollegioni et al. 2014). Besides morpho-agronomical characteristics, several biochemical (such as enzymes) and molecular markers have been established to identify walnut varieties and to estimate genetic diversity in walnut (Mahmoodi et al. 2013;Ali et al. 2016;Vischi et al. 2017). The identi cation of varieties based on phenotypic observations is slow, di cult, produces uncertain results because of environmentally induced variability, and it may even be impossible for juvenile or recently grafted plants (Pourkhaloee et al. 2017;Nadeem et al. 2018). Random ampli ed polymorphic DNA (RAPD) (Ahmed et al. 2012;Pop et al. 2013;Salieh et al. 2013), Restriction fragment length polymorphism (RFLP) (Fjellstrom et al. 1994), Inter-Simple Sequence Repeat (ISSR) (Jia et al. 2011;Grouh et al. 2011), simple sequence repeats or microsatellites (SSR) (Pop et al. 2013;Victory et al. 2006;Wang et al. 2008;Topcu et al. 2015;Vahdati et al. 2015), and ampli ed fragment length polymorphism (AFLP) (Kafkas et al. 2005, Bayazit et al. 2007) are the most common molecular markers which have been applied in walnut. DNA ngerprinting is commonly used to con rm the delity of varieties. Short tandem repeats DNA sequences known as microsatellites lie between conserved sequences and are scattered throughout the eukaryotic genomes (Bernard et al. 2018). Inter simple sequence repeat (ISSR) are also used to amplify DNA found between microsatellites, by using the PCR reaction with a repetitive sequence anchored by arbitrary or degenerate nucleotides. As a result, a large number of bands useful for ngerprinting purposes are obtained. ISSRs have been applied successfully in population genetic studies for a variety of organisms, including clonal plants (Wang et al. 2008).
SSR and ISSR are two of the main markers used in molecular studies because they are highly reproducible, show high level of polymorphism and are based on PCR procedure (Noor Shah et al. 2018). SSR and ISSR molecular markers are especially used in plant breeding for "DNA ngerprinting" important clones and for cultivar identi cation (Powell et al. 1996;Pollegioni et al. 2006;Dangl et al. 2005;Naja et al. 2014). Microsatellite also appears to be the method of choice for ngerprinting varieties and germplasms (McGranahan 2007). SSRs and ISSRs permit rapid and accurate identi cation of errors such as misidenti ed trees before they are delivered to growers (Oraguzie et al. 2005). DNA ngerprinting by SSRs and ISSRs is a rapid method for quality control, con rmation of variety (true-to-type) and veri cation of identity (Matsumoto and Fukui 1996). DNA ngerprints also have great importance in certi cation Our purpose was to use SSR and ISSR molecular markers to certify and prove the identities of clonal ramets in a walnut orchard by comparison with source plants. The identi cation of commercial Persian walnut varieties and/or proving their similarity using an SSR-based DNA and ISSR ngerprint and the publication of the alleles found in each cultivar at SSR and ISSR loci can be used to help verify true-to-type and to certify nursery stocks for commercial sale.

Plant material
In the present study, 11 varieties of Persian walnut (Juglans regia) growing at the mother orchard of Seed and Plant Improvement Research Institute in Kamalshahr, Karaj, Iran (35°40'N 51°19'E) were used as source of scion wood (Fig.   1). Twenty eight grafted ramets, propagated using the scion woods taken from the same 11 varieties described above, were grown at the walnut research orchard of Department of Horticulture, College of Aburaihan, University of Tehran, were used for SSR analysis (Table 1). Young leaves of mother plants and their clones were sampled from each tree in spring and stored at -80 °C until used for DNA extraction.
DNA extraction DNA extraction from collected samples were performed using the method of Doyle and Doyle (Doyle and Doyle 1990) with some modi cations as follows: about 100 mg of freeze-dried young leaves was ground in liquid nitrogen and leaves powder poured into a 1.5 mL tube; then 700 µL of preheated (65 °C) 2X CTAB buffer (2% CTAB, 2% PVP, 0.2% βmercapto ethanol, 1.4 M NaCl (pH=8.0), 100 mM Tris-HCl (pH= 8.0), 20 mM Na 2 EDTA (pH= 8.0) was added, the nal mix was incubated at 65 °C for 40 min followed by mixing with an equal volume of chloroform/isoamyl alcohol (24:1) and centrifugation at 13,000 g for 15 min. The resulting aqueous phase was taken and mixed twice with chloroform to remove contaminants. After centrifugation, the upper (aqueous) phase layer was taken and mixed with cold isopropanol for DNA precipitation. After centrifugation at 13,000 g for 10 min, nucleic acid precipitate was washed by 70 % ethanol for 0.5 hour, dried at room temperature and the resulting DNA pellet was dissolved in 50 µL TE buffer (pH=8). The DNA samples were stored at -20 °C. DNA was quali ed in 2 % agarose gel in TAE buffer, stained with ethidium bromide (10 mg/ml) against phage DNA standard and visualized under UV light (Geldoc, UVP CO, USA). DNA quantity was determined using a spectrophotometer (Perkin Elmer Co.).
For each locus and sample, PCR reactions contained 60 ng genomic DNA, 2 mM MgCl2, 0.2 mM dNTP, 1X PCR buffer, 1 unit of Taq DNA polymerase (CinnaGen Co.), 50 ng µL ¹ of each primers (forward and reverse primers) in a nal 20 µL volume. Microsatellites were ampli ed using a BioRad Gradient thermal cycler (Bio-Rad Laboratories, Inc., CA). PCR pro les were as follows: initial denaturation step at 94°C for 5 min, 35 ampli cation cycles of 45s at 94°C, 1 min at 52°C and 1 min at 72°C with a nal extension of 10 min at 72°C. To prepare samples for loading, twenty microliters of denaturing dye solution (95% formamide, 0.5% bromophenol blue, and 0.5% xylene cyanol) was added to 15 µL of the remaining reaction product. The diluted samples were denatured by heating to 96°C for 3 min, and then immediately chilled on ice, 7 or 8 µL of the diluted mixture was loaded in every lane of warmed 6% polyacrylamide gels using a 49 well paper comb onto a BioRad sequencer, under denaturing conditions (7M urea, 1X TBE buffer (90 mM Tris -borate, pH 8.3, and 2 mM EDTA pH=8.0). DNA ladder (Fermentas) as allele size standard was run in every gel to determinate the PCR products size. Gels were stained with silver solution (Bassam and Gresshoff 2007) (Fig. 2).
Photographs were taken of each gel using a Polaroid camera (DS-34). A 100 base-pair ladder was used to estimate the molecular size of bands. ISSR reactions were repeated at least 3 times and bands that were not consistent in all 3 reactions were discarded. Although a very high number of large amplicons were observed, the proximity between bands or their low intensity made scoring di cult. Highly reproducible bands were 300 to 1500 bp fragments.
Candidate amplicons within the selected size-range were discarded if they were close in size to other fragments or unclear when they were presented in mother plants and their clones, simultaneously. ISSR bands were scored as absent (0) or present (1), assuming that each band position corresponded to 1 'locus' with 2 'alleles' (1, 0).

Data analysis
The SSR and ISSR alleles of each clone were scored as present (1) or absent (0). Nei's genetic distance was calculated among all samples on the basis of allele frequency data. Two similarity matrices were generated based on the proportion of shared ampli cation fragments separately for SSR and ISSR data (Nei and Li 1979). Number (Na) of alleles, effective number of alleles (Ne) per locus and observed heterozygosity (Ho) were calculated for SSR loci using Power Marker V3.2.5. The polymorphism information content of primers was calculated as: where p i was the frequency of the i th genotype of each SSR locus.
Cluster analysis was carried out using the Jaccard similarity test and an UPGMA clustering method (Durvasula and Rao, 2018) using the software Darwin V6.0.21. The data from the polymorphic markers were subjected to population structure analysis based on the admixture model-based clustering method in the software package STRUCTURE 2.3.4 (Falush et al., 2007). This model was run by varying the number of clusters (k) from 1 to 11 and 4 iteration for each K.
A burn-in period of 10,000 and Markov Chain Monte Carlo (MCMC) replications of 50,000 after each burn-in was used. The Evanno method was used to estimate the ∆K which gives the best estimate for an optimum number of clusters (Evanno et al., 2005). The admixture model was repeated for the best K with a burn-in period of 50,000 and an MCMC of 100,000 at 1 iteration.
To determine the relationships among 39 samples, principal component analysis (PCA) was performed using DARwin V6.0.21 software.

Results And Discussion
Page 5/22

SSR marker analysis
The identi cation of varieties within and between walnut orchards is important for better management of genetic resources in the collections. In this study, the ampli cation of seventeen SSR loci generated 61 total fragments (alleles) from 28 individual clonal plants of eleven Persian walnut varieties. The least polymorphic locus in our study was WGA004 (2 alleles per locus) and the most polymorphic (5 alleles) was at WGA276. The average number of alleles per locus was 3.5 ( Table 2). The most common allele (91%) was at WGA004 (230 bp) ( Table 2). The size range of amplicons was assayed between 162 (WGA069) and 277 bp (WGA331) ( Table 2). Polymorphism at SSR loci in Persian walnut is population-dependent, but some loci have consistently been shown to be more or less polymorphic. Dangl et al. (2005) studied the genetic diversity of 47 J. regia genotypes using 14 SSR primer pairs and identi ed 73 alleles with an average of 2.5 alleles per locus. Li et al. (2007) observed only three alleles at WGA001 and seven at WGA005 and WGA071. Foroni et al. (2006) analyzed 12 SSR loci and reported fewest alleles (N = 3) at WGA001 and WGA118, and the maximum number of alleles (N = 8) at WGA071. In the study of Bernard et al. (2018) among the 217 J. regia accessions, the number of alleles per locus ranges from 2 to 17 with an average of 8.92 and the number of effective alleles ranges from 1.18 to 4.61 with an average of 2.84.
Level of polymorphism strongly determines how useful an SSR locus will be; a small number of highly polymorphic microsatellite loci can often provide enough information to discriminate among genotypes. Foroni et al. (2005) used six SSR primer pairs to study the Italian walnut landraces in Sorento. Li et al. (2007) showed only ve microsatellite markers were needed to obtain enough polymorphisms for the discrimination of 29 rootstocks of peach. In their study, the total number of observed alleles was 33, with an average of 5.5 alleles per locus. Vaiman et al. (1994) stated that if the PIC is higher than 0.5, then that locus has a high polymorphism level in the population examined; if it is between 0.25-0.5 the locus was considered moderately polymorphic, and if it is less than 0.25 its polymorphism is low. In this study PIC values ranged from 0.43 (WGA032) to 0.08 (WGA004) with a mean value of 0.29 (Table 3). Dangl et al. (2005) reported the highest PIC for WGA202 and WGA321 (0.74).
Observed heterozygosity in our study populations was moderate and nearly identical (H O = 0.56), a value somewhat lower than in similar studies. The per-locus observed heterozygosity ranged from 0.18 in WGA004 to 0.90 in WGA032 (Table 3). Foroni et al. (2005) found that in their population the average observed heterozygosity was 0.68, and the locus with the highest observed heterozygosity was WGA071 (0.81). In the study of Ebrahimi et al. (2011), the average value of observed heterozygosity across loci was 0.72. The per locus observed heterozygosity (Ho) ranged from 0.10 to 0.68 with a mean value of 0.47 in the study of Bernard et al. (2018).
The sizes of SSR alleles at a locus can vary across populations and can vary among genotypes even for clonal varieties (Nybom et al. 2014). Allele sizes we observed for some commercial varieties at some loci matched the sizes mentioned in studies by Dangl et al. (2005) and Foroni et al. (2005Foroni et al. ( , 2006. New alleles found in this research were allele d at WGA001 found in Iranian varieties, Z63 and K72, allele d in WGA071 found in the Iranian variety Z30, and allele e at WGA276 found in the Iranian variety Z53 and 'Pedro' (Table 2).
In this study, WGA001, WGA009, WGA069 and WGA202 were the most informative markers, while WGA004, WGA027, WGA032 and WGA376 did not have any polymorphism useful for discriminating among these samples. Using 13 polymorphic SSRs, we tabulated a DNA ngerprint chart of 11 maternal varieties (Table 5). Allele sizes for each of the ve replicates of each clone were largely consistent (Table 2), and some allelic combinations at a locus were unique to the varieties we tested (Table 5). The bold, underlined entries were unique allele combinations that can be used to identify clones (Table 5). Except for 'Chandler', each cultivar could be distinguished using a combination of only two SSR loci. These 13 SSR markers were powerful enough to produce a unique DNA ngerprint for each of the 11 maternal varieties, so they can be used to certify clonal identity with respect to varieties in this study. According to the UPOV guidelines (UPOV International Union for the Protection of New Varieties of Plants 2010) for DNA ngerprinting using molecular markers for the protection of new plant varieties, only markers with distinct PCR bands, high reproducibility, and reasonable polymorphism can be selected to use for this purpose. SSR primer sets for cultivar identi cation have been identi ed for many economically important plants, such as Camellia (Chen et al. 2016 ISSR analysis of 15 markers produced 38 scorable fragments, of which 33 were polymorphic ( Table 6). The number of alleles observed at each ISSR locus varied from 1 (AGA (AC) 7) to 10 (CAC (TGT) 5) with an average of 3.78 alleles per locus. The size ranges of the bands were between 100 to 1000 bp. The most proli c ISSR primer was CAC (TGT) 5 amplifying 10 polymorphic bands, while the least productive markers amplifying single monomorphic loci were AGA (AC) 7. ISSR markers have the potential to discriminate walnut varieties from each other because they generate a large number of products per primer. PIC values range from 0.11 (AGA (AC) 7) to 0.49 (CAC(TGT)5) with an average of 0.19.
Nei's genetic distance based on ISSR markers ranged from 0.15 to 0.74 with an average of 0.58 among all the walnut varieties in our study, which was lower than the ndings of Mahmoodi et al. (2012) who used ISSR analysis for the characterization and categorizing of walnut genotypes of Jammu and Kashmir. Effective number of alleles (Ne) and Nei's gene diversity were 1.48 and 0.29, respectively (Shah et al. 2019). Ghanbari et al. (2019) evaluated thirty one genotypes of walnut using 10 ISSR markers and identi ed a total of 26 polymorphic alleles, with an average of seven alleles per locus (Ghanbari et al. 2019). In this study, the lowest pair-wise dissimilarity (15%) was observed between 'Vina' and 'Franquette' while as highest pair-wise dissimilarity (74%) was observed between 'Serr' and Z30, which reinforced results of the SSR markers. The Mantel test showing no signi cant coe cient of correlation between SSR and ISSR distance matrices (r = 0.34, P = 0.05).

Cluster analysis based on combination of SSR and ISSR markers
To better visualize the relationships among the varieties, the data from both the primer types (SSR and ISSR) was pooled and a dendrogram was developed from this data. The dendrogram generated from the UPGMA cluster analysis showed considerable variation among all 11 varieties (Fig. 3). The optimum number of sub-populations K, which best explain the population structure of the accession was estimated to be 3 (K=3) using the Evanno method (Evanno et al., 2005). The population structure from the structural analysis is shown in Fig. 4, which depicts the three subpopulations composed of the 39 accession studied.
Cluster analysis revealed that within the rst group, three sub clusters were seen. In the rst sub cluster, Z63 and 'Lara' samples were closely grouped together. Z53 and Z30 and their ramets were grouped in the second and the third sub clusters of rst group, respectively. K72, 'Hartley', 'Serr' and their ramets were separated in the second group, with the exception of Serr.4, which was genetically distinct but similar to other 'Serr' ramets. The third group included ramets of 'Franquette' 'Chandler', 'Vina' and 'Pedro'. Vina1 was separated from other 'Vina' samples in this group (as an off-type); it was a distinct clade and was in fact genetically distinct from all other clones. Except for two samples that appeared to be off-types, a unique genotype was obtained for the ramets of each clone, and each ramet of each clone had the same genotype. These results con rm the rametʼs true-to-type except for the Vina.1 and Serr.4 variants, so Jaccard's coe cient between the ramets and the stock block clones was 1. The Jaccard's coe cient between Vina. 1 and Vina was 0.28 and between Serr. 4 and Serr was 0.56 (Table 4). Principal coordinate analyses (Fig 5) con rmed the results of the cluster analysis and revealed three well-separated clusters. All ramets of the same source genotypes were placed with their source clones, with the exception of Vina1 and Serr4. The reasons for misclassi cation may have included mislabeling of samples, failed grafts, and collecting the wrong scion wood from the mother orchard. This kind of mistake usually happens when grafters do not use right labeling protocols and/or do not use different painting colors to characterize each cultivar.

Conclusion
Our results demonstrate the use of microsatellites and Inter-simple Sequence Repeat markers for identi cation and veri cation of Persian walnut (Juglans regia) varieties and the testing of trueness to type of grafted ramets. Applied research of this type is essential to the nursery trade, which needs simple, low-cost methods for certifying stock and the prevention of errors in propagation. Our results also show the value of using microsatellites to verify genotype identi cation in breeding programs where propagation errors can result in misallocation of effort, poor data, and even failure to show progress.

Declarations
Disclaimer: Mention of trademark, proprietary product, or vendor does not constitute a guarantee or warranty of the product by the U.S. Dept. of Agriculture and does not imply its approval to the exclusion of other products or vendors that also may be suitable.

Data archiving statement
The authors con rm that the data supporting the ndings of this study are available within the article.

Funding information
Funding for this research was provided by the University of Tehran.
Compliance with ethical standards

Con ict of interest
The authors declare that they have no con ict of interest.