DNA fingerprinting of Japanese plum (Prunus salicina) cultivars based on microsatellite markers

Forty-seven Japanese plum (Prunus salicina) cultivars were genotyped with eight microsatellite markers, aiming at obtaining the DNA fingerprinting profiling, distinguishing and characterizing a representative set of Japanese plum cultivars. The eight SSR loci amplified 104 alleles (8 to 21 alleles per locus, mean 13). Polymorphism Information Content (PIC) ranged from 0.680 to 0.886 (mean 0.803). The observed heterozigozity (Ho) ranged from 0.529 to 0.915 (mean 0.770). Probability of Identity (I) of each locus ranged from 0.019 to 0.113 (mean 0.054). The combined Probability of Identity was 2.66 x 1011, and the Power of Exclusion of the eight loci was 99.99976%. 57 out of 104 alleles showed frequency lower than 0.05. These low allele frequencies contributed to raise the distinguishability of plum cultivars. These results will contribute, as excellent descriptors, to select parental for crossings, to perform early identification of segregating clones with potential to be cultivars, and to protect the cultivars.


INTRODUCTION
Japanese plum (Prunus salicina) is one of the two predominating species in large-scale commercial plum production. Varieties of this species have a wide range of adaptation from temperate regions to the subtropics and are the predominant fresh market type in America and Asia. The other species, the hexaploid European plum (Prunus domestica), is more adapted to cool temperate climates Weinberger 1996, Topp et al. 2012).
The term Japanese plum was originally applied to P. salicina, but now includes all the freshmarket plums developed by intercrossing various diploid species with the original one. Although it is native to China, this species was initially improved in Japan, and later, to a much greater extent, in the United States (Okie and Hancock 2008).
At the end of the 19th century, Luther Burbank crossed imported P. salicina with P. simonii and several American plum species. These cultivars formed the base for current cultivars. In 1996, eight of the top ten producing Californian cultivars had Luther Burbank cultivars in their ancestry (Okie and Ramming 1999). A consequence of this polyspecific breeding, associated with the natural outcrossing of Japanese plums, is the great variability among plum cultivars. Byrne (1990) found that the mean inbreeding and coancestry coefficients for plum were one half or less than those for peach. Currently, Japanese plum cultivars are a mixture of P. salicina and at least one other plum species.
Plum cultivar identification can be very difficult when relying upon morphological characteristics alone. DNAbased markers, and particularly microsatellites (or SSRs), are very useful tools for distinguishing cultivars since they directly reflect the genotype. In addition, there is a large number of potential polymorphic sequences available for distinct genetic studies. Moreover, microsatellite markers also showed to be a powerful tool for genetic characterization of plum varieties (Ahmad et al. 2004); thus, they can be used to solve cases of taxonomic synonyms, misidentification, and patent or protection issues. Since effective utilization of germplasm resources depends on accurate and unambiguous characterization, microsatellites can also help breeders in their breeding programs.
The objectives of this work were to perform the genetic characterization, and to establish a DNA fingerprinting of commercial plum cultivars. With the use of SSR markers, GHF Klabunde et al.
cultivars from the Prunus breeding program at EPAGRI-Videira Experimental Station could be discriminated from other cultivars.

Plant material and DNA isolation
The forty-seven Japanese plum cultivars (Table 1) selected for this study represent a wide genetic spectrum, containing several target genes, particularly to the plum breeding in south Brazil. Currently, some of these cultivars are commercially cultivated in south Brazil, such as Fortune and Laetitia cultivars. Other cultivars represent the genetic pool for the Japanese plum breeding, such as Chatard, Piamontesa and Carazinho, which are highly resistant to leaf scald (Dalbó et al. 2010). The Japanese plum collection is located at Epagri -Videira Experimental Station (Videira, SC, Brazil), lat 27° 00' 30" S, long 51° 09' 06" W, and alt 800 m asl. DNA isolation was carried out with a modified method of Doyle and Doyle (1990), as described in Vieira et al. (2005).

SSR markers
Eight genomic microsatellite markers chosen for the genetic characterization (Table 2) were originally developed by Cipriani et al. (1999), Dirlewanger et al. (2002), Yamamoto et al. (2002) and Mnejja et al. (2004). The choice was based on the fact that these markers are highly polymorphic, and because they amplified microsatellite sequences that are located in distinct linkage groups. CPSCT markers (Mnejja et al. 2004) were designed originally to Prunus salicina Lindell. The other microsatellites were designed to peach (Prunus persica [L.] Batsch) and apricot (Prunus armeniaca L.), and were used for being highly polymorphic in several Prunus species, such as almond (Prunus dulcis), sweet cherry (Prunus avium L.), apricot (Prunus armeniaca L.), and others (Mnejja et al. 2010). Each 5´ forward oligo was labelled with a fluorophore in order to enable the automatized genotyping (Table 2).

PCR conditions and genotyping
PCR reactions were performed in a volume of 15 µl containing 10 mMTris -HCl pH 8.8, 50 mMKCl, 1.5 mM of MgCl2, 0.2 mM of each dNTP, 0.3 µM of each primer, 15 ng of genomic DNA, and 1 U of Taq DNA polymerase (AmpliTaq Gold® Invitrogen). A BioRad C1000 Thermal Cycler was used to amplify the eight SSR loci with the following cycling profile: 94 ºC for 3 min, then 30 cycles of 94 ºC for 30 s; annealing temperature specific to each primer (see table 2 for each SSR temperature) for 30 s, and 72 ºC for 30 s; and a final extension step of 5 min at 72 ºC. After the reactions, all PCR products were diluted 20X in ultrapure water in order to be genotyped by capillary electrophoresis in a MegaBACE 1000 DNA Analysis System (GE Healthcare) DNA sequencer. Alleles were genotyped by comparison with ET 400-R size standard (GE Healthcare), using Fragment Profiler software version 1.2 (GE Healthcare).

Data analysis
The Observed heterozygosity (Ho), number of alleles per locus (A), allele frequencies, Polymorphism Information Content (PIC; being PICi = 1 -Ʃ Pi 2 , where Pi is the frequency of allele I band), Probability of Identity (I; being I = Ʃpi 4 + Ʃ (2pipj) 2 ), where p i and pj are the frequencies of the ith and jth alleles and i ≠ j) and (I unbiased = n 3 (2a 2 2 -a 4 ) -2n 2 (a 3 + 2a 2 ) + n(9a 2 + 2) -6/(n -1)(n -2)(n -3)) where n is the sample size, a i equals Ʃpj i and p j is the frequency of the jth allele (Paetkau and Strobeck 1994), and Power of Exclusion (Q) (Vandeputte 2012) were calculated using Cervus 3.0 software (Kalinowski et al. 2007). Alleles were considered rare when their frequencies were less than 0.05. A dendrogram was constructed via the unweighted pairgroup method with arithmetic means (UPGMA) (Sneath and Sokal 1973), using the Darwin 5.0 software (Perrier et al. 2003), and based on Pearson (r) similarity coefficient and cophenetic correlation coefficient.

RESULTS AND DISCUSSION
The eight SSR markers amplified a total of 104 alleles, showing 8 to 21 alleles per locus, with a mean of 13 alleles ( Table 3). The great majority of alleles (54.8%) were considered rare, with frequencies less than 0.05 ( Figure  1). Among these alleles, 27 only occurred once. On the other hand, only 11.5% of the alleles showed frequencies higher than 0.2 (Figure 1). This pattern was expected, since a diverse group of cultivars was genotyped (Table  4). Heterozygosity ranged from 0.529 to 0.915 per marker, with mean value of 0.770 (Table 3). UDP 97-402 marker presented the lower heterozygosity value due to the high presence of null alleles (13 of 47 cultivars, Table 4). Non amplified fragments were considered null alleles since the DNA amplified in other markers and the PCR conditions were optimized based on Cipriani et al. (1999). In addition,    The results also allowed validating some cases of parentage reported in the literature. Laetitia (Letícia in Brazil) is reported as being a descendant of the cultivar Golden King by open pollination. The present results are consistent with this hypothesis, since there is always one allele shared by these two cultivars at all analyzed loci. The same occurs to Laroda and Santa Rosa, which is compatible with the information that Laroda results from the cross Gaviota x Santa Rosa. It was also confirmed a case of synonymy, as reported by Okie and Weinberger (1996), between Sordum and Gran Sultan, since both have the same alleles at each tested locus. The cultivar Gran Sultan is originally from New Zealand, and Sordum is an important cultivar in Japan. In this case, the spelling Sordum, which became popular in Japan, is probably a corruption of Sultan.
The number of alleles per locus (8 to 21) can be considered  high, especially if compared with the closest species, such as peach (Aranzana et al. 2003, Aranzana et al. 2010) and apricot (Villanova et al. 2006), and it is slightly lower than that of grape, in which it was found 13 to 23 alleles ). The observed allele richness was also higher than that reported by Ahmad et al. (2004), who found 2-10 alleles per locus in a study involving 14 cultivars of plum, 7 of apricot, and 7 hybrids (plumcots and pluots). This may be due to the number and diversity of plum cultivars analyzed in this study, or greater variability of loci examined, since they have already been selected for that trait.
The high genetic variability detected allowed cultivar identification by using a relatively small number of loci. Thus, the results of the present study provide the basis for characterization and identification of plum cultivars, and build a database for identification of genetic accessions of this species. A similar situation occurs in grape, with a similar degree of genetic diversity, having one set of six loci proposed as a reference for identification of genotypes of this species ).
The dendrogram (Figure 2) based on the genetic similarity among the different accessions brought four main groups, being two of them formed solely by one cultivar, Santa Rosa and Coeur de Lion, respectively. The other two groups are divided in several other subgroups, according to genetic proximity. The high genetic similarity is expected between individuals of the same subgroup, once they share a common origin. However, although the detected molecular differences between the accessions were not of great magnitude, it is possible to conclude that the vast majority of analyzed accessions differ among themselves, and this feature is essential in the management of a Plum Active Germplasm Bank and for fingerprinting issues.