Genetic variability evaluation and cultivar identification of tetraploid annual ryegrass using SSR markers

Plant materials

The plant materials used in this study included six tetraploid major commercial cultivars of annual ryegrass which were registered in China (Table 1). Seeds of each cultivar were germinated in a petri dish (Biosharp™, Hefei, China, diameter= 90 mm) with two layers of moistened filter paper. The germination was performed in a growth chamber with a constant temperature of 25 °C. One hundred robust seedlings of each cultivar were selected to transplant into a sand-peat mixture (Panshi™, He’nan, China). All the seedlings were grown in a greenhouse with a 12 h photoperiod at average temperatures of 25/20 °C (day/night) for two weeks. Plants were watered every two days and not fertilized during the experiment.

DNA extraction

For the DNA extraction, fresh young leaves were collected from each individual and a total of 10 independent bulks were constructed for each cultivar. For each bulk, samples of 10 randomly selected leaves were mixed. Genomic DNA was extracted using the genomic DNA extraction kit (Tiangen^®, Beijing, China) according to the manufacturer’s protocol, then the quality and concentration of the DNA were measured on 0.8% (w/v) agarose gels and NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies Inc., Rockland, DE, USA), respectively. DNA samples were diluted to 20 ng µl⁻¹ and stored at 4 °C for PCR amplification.

SSR amplification

A total of 29 pairs of SSR primers selected from Hirata et al. (2006) were used in this study (Table 2, Supplemental Information 1). PCR amplifications were carried out in a volume of 15 µl containing 1.5 µl (20 ng µl⁻¹) genomic DNA, 7.5 µl PCR-Mixture (10 × reaction buffer, 2.0 mM Mg²⁺, 0.6 mM of each dNTP, Tiangen^®, Beijing, China), 1.2 µl (10 pmol µl⁻¹) forward and reverse primer mixture, 0.3 µl of Taq DNA polymerase, and 4.5 µl of ddH₂O. PCR amplification reaction was performed using touchdown mode with the following program: 5 min at 94 °C for 1 cycle, followed by 10 cycles at 94 °C for 30 s, varying annealing for 45 s according to annealing temperature (Tm) of different primers (Table 2) with each cycle decreased by 0.5 °C, and 72 °C for 1 min, and then 30 cycles at 94 °C for 30 s, annealing temperature for 45 s, and 72 °C for 1min, final extension at 72 °C for 7 min, and then stored at 4 °C. The amplified fragments were separated on 6% denatured polyacrylamide gel electrophoresis using 200V pre-electrophoresis for 30 min, and then 400V electrophoresis for 2 h. After electrophoresis, the gel was stained by AgNO₃ solution and photographed by camera.

Data analysis

The amplified fragments of all SSR primers were scored as band presence (1) or absence (0) to form binary matrix, and each of them was treated as an independent character regardless of its intensity. The total number of bands (TNB), number of polymorphic bands (NPB), and percentage of polymorphic bands (PPB) were calculated. The discriminatory power was evaluated by polymorphic information content (PIC) calculated as ${\mathrm{PIC}}_i=2f_i\left(1-f_i\right)$, where PIC_i is the polymorphic information content of marker ‘i’, f_i is the frequency of the amplified allele (band present), and 1 − f_i is the frequency of the null allele (Roldan–Ruiz et al., 2000).

The analysis of molecular variance (AMOVA) was used to estimate the genetic variation within and among cultivars in WIN AMOVA software (Excoffier, Smouse & Quattro, 1992). The genetic similarities among 10 bulks within each cultivar were calculated with NTSYS-pc2.10 software. The principal coordinate analysis (PCoA) and the un-weighted pair group method with arithmetic mean (UPGMA) dendrogram were constructed by NTSYS-pc 2.10 software and MEGA v6.0 (Tamura et al., 2013), respectively.

In order to define the optimal threshold of bands frequencies for ryegrass cultivar identification, the comparison analysis of difference thresholds for filtering statistical bands was conducted. The bands appeared 10 times (100% frequencies), nine times (90% frequencies), eight times (80% frequencies), seven times (70% frequencies), six times (60% frequencies), five times (50% frequencies), four times (40% frequencies), three times (30% frequencies), two times (20% frequencies) and one time (10% frequencies) within 10 bulks of each cultivar and were respectively recorded to form a new data matrix. The Dice’s similarity coefficients for estimating genetic similarities between cultivars were determined by using NTSYS-pc 2.10 software, and the UPGMA dendrogram were constructed for each filtering statistical data matrix. Each pair of primers was integrated with a single fingerprint pattern for evaluating the discriminatory power.

Genetic variation and clustering analysis of the six tetraploid annual ryegrass cultivars

A total of 242 reliable bands were obtained from 29 SSR primer pairs, with an average of 8.3 bands for each primer ranging from 3 (LMgSSR16-01E) to 14 (LMgSSR04-05B) (Fig. 1, Table 2). Among them, 227 bands (93.8%) were polymorphic and the percentage of polymorphic bands (PPB) per primer pair ranged from 57.1% to 100%. Meanwhile, the polymorphic information content (PIC) values for each primer ranged from 0.170 (LMgSSR17-04E) to 0.421 (LMgSSR00-04A), with an average of 0.304, demonstrating potentially sufficient discriminatory capacity. The AMOVA analysis was conducted to investigate the genetic variation within and among cultivars. The results revealed that the majority of the total variation was due to within-cultivars (81.99%), while variance among cultivars was only responsible for 18.01% (Table 3).

The genetic similarity (GS) analysis among 10 bulks of each cultivar showed that the variation within cultivars was different and the GS coefficient value ranged from 0.639 (relative high variation within a cultivar) for the cultivar ‘Angus 1’ to 0.723 (relative low variation within cultivar) for the cultivar ‘Aderenalin’ (Supplemental Information 2). The Un-weighted Pair Group Method with Arithmetic Mean (UPGMA) dendrogram based on GS data showed that all six cultivars with 10 bulks were obviously distinguishable (Fig. 2). Among them, the cultivars ‘Double Barrel’ and ‘Abundant’ had closest family relationship with two bulks of ‘Abundant’ clustered with ‘Double Barrel’ group, followed by ‘Tetragold’, ‘Changjiang No.2’ and ‘Angus 1’. The cultivar ‘Aderenalin’ was divided from the other cultivars, indicating that the highest genetic distance and different parental ancestry compared to others. One bulk of ‘Changjiang No.2’ and one bulk of ‘Tetragold’ separated from other bulks of their respective cultivars to form a separate branch, respectively.

The principal coordinate analysis (PCoA) provided better understanding of the relationships among the tested annual ryegrass cultivars. The results showed that the total variation could well be explained by the first two principal axes, with explanation rate of 8.57% (PC1) and 6.05% (PC2), respectively. The results of PCoA were consistent with the cluster analysis that the bulks of cultivar ‘Aderenalin’ were significantly distinguished by PC2 (group 1). ‘Changjiang No.2’ could be clearly separated by PC1 (group 2), while UPGMA dendrogram based on GS data did not distinguish it. The other cultivars were distributed around the origin of coordinates forming a group 3 (Fig. 3).

Cultivar identification of tetraploid annual ryegrass

Although multi-bulk strategy could increase the frequency of rare alleles, the identification of a suitable threshold for filtering scored bands is a guarantee for stable results. In this study, the comparison analysis of difference thresholds among 10 independent bulks of each cultivar was conducted, and the results showed that the identification ability varied among SSR primer pairs and different filtering strategies (Table 4). Based on difference filtering threshold, the mean of directly distinguishable number of studied cultivars ranged from 1.4 (100% frequencies) to 4.5 (30% frequencies), and comparable identification abilities were observed when scored bands appeared two, three, or four times among 10 bulks (the number of cultivars could be directly distinguished was above four).

Among the 29 SSR primer pairs, 9 of them could effectively identify six cultivars when the bands appeared frequency at 60%, 10 at 50%, 12 at 40%, 14 at 30% and 16 at 20%, whereas just two of them were available for directly identifying six cultivars at 70%, 80% and 90% (Table 4). In addition, the identification ability could always be improved by a combination of different primer pairs. For example, at the threshold of 40%, all six cultivars could be identified by a combination of the following primers (primer 02-05G could identify ‘Angus 1’, ‘Changjiang No.2’, ‘Aderenalin’ and ‘Tetragold’, while primer 07-07G could identify ‘Angus 1’, ‘Double Barrel’, ‘Aderenalin’ and ‘Abundant’).

Suggestions for optimum multi-bulk strategy

To validate the effectiveness of multi-bulk strategy results, clustering analysis of binary scoring with difference thresholds was conducted. The UPGMA dendrogram showed that difference filtering strategies produced varied identification ability, while more stable results could be achieved with higher band presence frequency (Fig. 4). Interestingly, the two most related cultivars were the ‘Abundant’ and ‘Tetragold’ when the bands appeared at a frequency of 10% with GS coefficient value of 0.90, ‘Abundant’ and ‘Angus 1’ at 20% with GS coefficient value of 0.85, ‘Abundant’ and ‘Tetragold’ at 30% with GS coefficient value of 0.80, and were stably ‘Abundant’ and ‘Double Barrel’ after the bands appeared frequency above 40% with GS coefficient value ranging from 0.76 to 0.90. Moreover, the results showed that ‘Aderenalin’, introduced from Germany, was the most genetically distant cultivar from others, except when DNA bands appeared at the frequency of 50% and 90%, ‘Changjiang No.2’ was the most genetically distant one. ‘Changjiang No.2’ is a cultivar hybridized between ‘Aubade’, introduced from USA, and ‘Ganxuan No.1’, bred from diploid breeding line by chromosome doubling and radiation mutation by mass selection in China.

Although those cultivars with limited available pedigree information could be difficult to assess in regard to phylogenetic relationships, the SSR markers were sufficient for distinguishing the cultivars. In most filtering situations, the UPGMA dendrogram (Fig. 4) showed that all six examined cultivars were grouped into three clusters with ‘Aderenalin’ was clearly distinguished from the others and followed by ‘Changjiang No.2’ which was consistent with previous cluster (Fig. 2) and PCoA (Fig. 3) based on original data binary.

Taking into account the ability to differentiate the cultivars (the discriminating power) and the bands stability (the consistency of DNA fingerprints), the results suggested that DNA bands presence frequency of 40% could be used as a reliable threshold when using multi-bulk strategy for cultivar identification. In this study, 12 of 29 SSR primer pairs (00-04A, 02-06G, 02-08C, 03-05A, 04-05B, 10-09E, 12-01A, 13-02H, 13-12D, 14-06F, 15-01C and 17-10D) were available for directly identifying six cultivars and ‘Abundant’ and ‘Double Barrel’, introduced from the USA had the closest genetic relationships, followed by ‘Angus 1’ introduced from the USA, ‘Tetragold’ introduced from the USA, ‘Changjiang No.2’ and ‘Aderenalin’ introduced from Germany.