Characterization of SSR genomic abundance and identification of SSR markers for population genetics in Chinese jujube (Ziziphus jujuba Mill.)

SSR identification and primer design

The final assembly of Chinese jujube genome spans 437.65 Mb (98.6% of the estimated size) with 321.45 Mb anchored to the 12 pseudo-chromosomes (Liu et al., 2014b). The 12 pseudo-chromosomes sequences were downloaded from GenBank (accession number: CM003114– CM003125) and scanned using MSDB v2.4 software (http://msdb.biosv.com) to identify all perfect microsatellites (Weber, 1990). The identification criteria used for mono-, di-, tri-, tetra-, penta- and hexanucleotide repeats were default in MSDB, with a minimum of 12, 7, 5, 4, 4 and 4 repeats, respectively. After analyzing each chromosome independently, all 12 chromosomes were analyzed together. Genome sequences of Arabidopsis thaliana were downloaded from GenBank (accession number: CP002684– CP002688) and comparatively analyzed against the jujube genome.

After searching the complete jujube genome, primers were designed in PRIMER v5.0 (Clarke & Gorley, 2001) using loci with repeat numbers greater than 25, for possible high polymorphism, and less than 100 for fitting with the scoring size standards. Microsatellites with mononucleotide repeats were removed when primer were designed. All primers were designed according to the following parameters: (1) product size from 100 to 400 bp and (2) primer size from 18 to 22 bp with an optimum size of 20 bp. All primers were synthesized by Tsingke Biological Technology Co. (Beijing, China). After screening using a series of criteria, parts of forward primer 5′ ends were labeled with a fluorescent dye (FAM or HEX) from Tsingke Biological Technology Co.

Plant materials and genomic DNA isolation

In this study, we tested the amplification and polymorphism of all primers using three populations: two populations (17 individuals each) of jujube cultivars ‘Bianhesuan’ and ‘Huizao’ and one population (14 individuals) of Z. acidojujuba C.Y. Cheng et M.J. Liu, the ancestor of Chinese jujube (Liu & Wang, 2009). We treated all plants of a cultivar collected from its major planting regions as a population. In addition, 10 main jujube cultivars were obtained from the Xinzheng Jujube Research Institute, Henan, China, to test the amplification feasibility of SSR markers in different cultivars. One individual was collected for each cultivar. Sample information is provided in Table S1. Total genomic DNA was extracted from dried leaves with a modified CTAB method (Doyle & Doyle, 1987), quantified using a NanoDrop 2000c spectrophotometer (Thermo Scientific, Waltham, MA, USA), and checked for quality by gel electrophoresis.

Polymerase chain reaction (PCR) and fragment analysis

PCR amplifications were carried out in 15-µl reaction volumes containing 10–20 ng template DNA, 1× PCR Buffer, 1.5 mM MgCl₂, 0.4 µM of each dNTP, 0.2 µM of each primer and 1 unit of Taq DNA polymerase (Takara, Dalian, China). The PCR cycling profile consisted of an initial step of 95 °C for 5 min, followed by 35 cycles of 95 °C for 45 s, 50–55 °C (annealing temperature of each primer) for 45 s and 72 °C for 30 s, with a final extension step of 72 °C for 7 min. The PCR products were subsequently detected using a 3730XL Genetic Analyzer Sequencer (Applied Biosystems, Foster City, California). Allele sizing was performed in GENEMAPPER v4.0 (Applied Biosystems) by comparing alleles with a GeneScan-500LIZ size standard (Applied Biosystems).

Analysis of genetic diversity

The presence of null alleles, scoring errors and large allele dropout was checked using MICROCHECKER version 2.2.3 (Van Oosterhout et al., 2004).The null allele frequency (r), inbreeding coefficient (Fis) and linkage disequilibrium (LD) between all pairs of polymorphic loci were calculated using program GENEPOP 4.0 (Raymond & Rousset, 1995; Rousset, 2008). The LD was tested with 10,000 permutations. The number of alleles per locus (N_a), observed (H_o) and expected heterozygosity (H_e), deviations from Hardy-Weinberg equilibrium (HWE) were calculated using ARLEQUIN version 3.5 (Excoffier & Lischer, 2010). The HWE was tested with 1,000,000 steps in Markov chain.

Characterization of ssr distribution

The frequency and density of SSRs on each of the 12 jujube chromosomes ranged from 818.29 (chromosome 5) to 920.85 (chromosome 12) loci/Mb and from 16,035.53 (chromosome 5) to 17,750.21 (chromosome 12) bp/Mb, respectively. Analysis of 321.56 Mb of sequence data from the assembled jujube genome uncovered 280,596 SSRs covering 5.37 Mb. These SSRs included 142,795 mono-, 87,362 di-, 29,900 tri-, 15,827 tetra-, 3,237 penta- and 1,475 hexanucleotide repeats, which respectively corresponded to 46.88%, 35.12%, 10.87%, 5.22%, 1.27% and 0.73% of total SSRs. There were 373 SSRs with more than 100 repeat motifs in the jujube genome, the longest of which comprised 2,446 AT repeats. Analysis of 119.15 Mb of the A. thaliana genome revealed only two SSRs with more than 100 repeat motifs (201 AAG and 117 TCT). Average SSR sizes in jujube and A. thaliana genomes were 19.14 and 17.40 bp, respectively. Jujube genome SSR frequency and density were 872.60 loci/Mb and 16,700.42 bp/Mb, respectively, with corresponding values of 221.78 loci/Mb and 3,858.19 bp/Mb in A. thaliana.

SSR frequency and density varied with motif length. As motif length increased (from mono- to hexanucleotide repeats), the frequency of repeats decreased from 444.06 to 4.59 loci/Mb and density decreased from 7,829.71 to 121.64 bp/Mb (Fig. 1). The 10 most frequent motif types in the jujube genome were two mononucleotide (A/T, G/C), three dinucleotide (AT/TA, AG/CT, GT/AC), four trinucleotide (ATT/AAT, AAG/CTT, ATC/GAT, AAC/GTT) and one tetranucleotide (ATTT/AAAT) repeats (Table 1). When calculating the frequency of each motif type, we considered all possible types in a repeat sequence, such as two potential types (AG and GA) for the AG motif repeat and three potential types (ATT, TTA and TAT) for the ATT motif repeat. As for A. thaliana, no tetranucleotide repeats were observed among the 10 most frequent motif types. The average length, frequency and density of the 10 most frequent microsatellite motif types are provided in Table 1.

Primer design and evaluation

Taking no account of mononucleotide repeats, a total of 664 SSRs with repeat numbers greater than 25 and less than 100 were detected in jujube. We randomly selected 100 SSRs from the 664 loci and successfully designed 70 primer pairs. Screened in 2% agarose electrophoresis, 42 of the 70 primer pairs produced clear, unique amplification products of the expected size. All the 42 SSR markers showed their ability to detect polymorphisms across jujube cultivars and wild species. A set of 20 SSR markers that amplified the most easily scorable fragment polymorphisms was chosen to label with a fluorescent dye and evaluate polymorphism in the three populations. Characteristics of these 20 SSR loci are listed in Table 2. All 20 primer pairs were also amplified successfully in the 10 jujube cultivars. Among the 20 loci, N_a ranged from 6 to 21, with an average of 12.8 alleles per locus (Table 2). The mean value of r among the three populations ranged from 0 to 0.261. Loci Juju18 presented too many null alleles (r > 0.15) in all populations and was eliminated from further analysis. Average H_o and H_e across the three populations were 0.909 and 0.701, respectively. The N_a had the highest mean values (8.5) in population 1 (Z. acidojujuba) among the three populations (Table 3). Population 2 (‘Bianhesuan’) has the highest H_o (0.967) but lowest H_e (0.578). The H_o and H_e were zero at locus Juju64 in population 2. Of the 20 loci, 16 (all except Juju18, Juju57, Juju63, Juju64 and Juju68) showed significant deviation from HWE in two cultivar populations (Table 3). None loci showed significant deviation from HWE in population 1. The Fis of the three populations were −0.066, −0.490 and −0.305, respectively. Three locus pairs (Juju11 & Juju23, Juju11 & Juju25 and Juju25 & Juju42) showed signficant LD (P < 0.001) across the three populations.

Characterization of SSR abundance

The characterization of SSRs in our study was different from that reported by Xiao et al. (2015). This difference was caused by two reasons. One was that their study was based on scaffold sequences and analysis 396.18 Mb of sequence data (Xiao et al., 2015) and our study was based on assembled 12 chromosomes and analysis 321.56 Mb of sequence data. The other was the different identification criteria of defining SSR. Their identification criteria used for mono-, di-, tri-, tetra-, penta- and hexanucleotide repeats were a minimum of 10, 5, 5, 5, 5 and 5, respectively (Xiao et al., 2015); however, ours were a minimum of 12, 7, 5, 4, 4 and 4. We used stricter criteria for mono- and dinucleotide repeats and therefore got less SSR loci. Although the results of two studies were different, they got consistent conclusions.

We calculated the frequencies of perfect SSRs composed of 1–6-bp long motifs in jujube and A. thaliana. The abundance of certain repeat types differed between the two plants. Jujube had a much higher SSR density than A. thaliana, as well as mulberry, peach, apple, pear, grape, strawberry and Prunus mume (Xiao et al., 2015). SSRs containing A and T (A/T, AT/TA and ATT/AAT) were dominant in the jujube genome, consistent with its high AT content (66.59%, Liu et al., 2014b). High proportions of (A+ T)-rich repeats (Xiao et al., 2015) are also present in A. thaliana and other plants (Lagercrantz, Ellegren & Andersson, 1993). In protein-coding regions, the high proportion of (A+ T)-rich repeats is due to the poly(A/T) tails of densely scattered retroposed sequences and processed pseudogenes (Tóth, Gáspári & Jurka, 2000).

Microsatellites are associated with nonrepetitive DNA and have their highest frequencies in transcribed regions (Morgante, Hanafey & Powell, 2002). Besides their utility as genetic markers, SSRs have important developmental, gene regulatory and evolutionary functions (Lawson & Zhang, 2006). The function of the high-density SSRs in the complex jujube genome (Liu et al., 2014b) thus warrants further study.

The frequency of repeats in jujube decreased with increasing motif length, similar to sorghum, rice, Medicago, Populus, Brachypodium and Brassica oleracea (Iniguez-Luy, Voort & Osborn, 2008; Sonah et al., 2011). Mononucleotide repeats were the most abundant repeats. One of the 10 most abundant motif types in jujube were tetranucleotide motifs (ATTT/AAAT) which was not found in A. thaliana. This difference may be ascribed to the complexity of the jujube genome and its high SSR density, as well as its high AT content (66.59%, Liu et al., 2014b).

Primer evaluation

More alleles were detected in population 1 than in populations 2 and 3. This difference may be partly explained by the fact that population 1 consisted of a wild species whereas populations 2 and 3 were cultivars. It is relevant to the lower Fis in population 1 than in populations 2 and 3. Wild species often have higher genetic polymorphism levels than cultivars, such as peanut (Moretzsohn et al., 2004), soybean (Li et al., 2010), apple (Zhang et al., 2012). Using nine microsatellites, the average H_o and H_e of 31 Z. acidojujuba populations was 0.679 and 0.659, respectively (Zhang et al., 2015). Comparing the Z. acidojujuba population in our study (from Luoyang) with their population from Luoyang (EYYC), our markers identified higher N_a (8.5 vs. 3.9) and higher H_o and H_e (0.901 and 0.844 vs. 0.683 and 0.625, respectively) (Zhang et al., 2015). The higher values hinted that our marks could uncover subtler genetic structure. Locus Juju64 uncovered only one allele in population 2, causing H_o and H_e to be zero, and also uncovered least alleles in other two populations, thus revealing the low polymorphism of the locus.

Out of the 70 screened primer pairs, 42 (60.0%) produced clear, unique amplification products and 20 (28.6%) displayed high levels of polymorphism. Loci Juju18 presented too many null alleles thus was eliminated. After testing in three populations, all the remained 19 polymorphic SSRs were in HWE in at least one population. For our populations, especially population 2 and 3, many loci were not in HWE because cultivars are not in a random mating system. Taking LD into consideration, the removal of one locus (i.e., Juju11 and Juju25) from each linked pair would give 17 unlinked loci in HWE.

In conclusion, we developed 17 validated microsatellite primer pairs with applicability to jujube population genetics. These polymorphic markers will facilitate the study of jujube genetic structure and gene flow and aid investigations of the evolutionary history of this important fruit crop.