The DNA methylation level against the background of the genome size and t-heterochromatin content in some species of the genus Secale L

Plant material

The following taxa of Secale were included in the study: Secale cereale L. cv. Dankowskie Zielonkawe, Secale cereale ssp. afghanicum, Secale cereale ssp. segetale, Secale strictum, Secale strictum ssp. africanum, Secale strictum ssp. kuprijanovii, Secale sylvestre and Secale vavilovii.

Caryopses of all taxa mentioned above were obtained from Botanical Garden of Polish Academy of Science (Warsaw, Poland).

DNA isolation

Total genomic DNA was isolated from 7-day-old seedlings using the DNeasy Plant Mini Kit (Qiagen, Valencia, USA) according to the manufacturer’s instruction. Each DNA extract was prepared from 30 randomly selected seedlings. The quality and quantity of the DNA samples was assessed by 0.8% agarose electrophoresis in 1×TBE buffer and measurements done using Nanodrop 2000c spectrophotometer (Thermo Scientific, Vilnius, Lithuania).

MSAP

500 ng of genomic DNA was digested with 1 µl of FastDigest^® EcoRI (Fermentas, Vilnius, Lithuania) in a total volume of 50 µl for 20 min at 37°C. After EcoRI inactivation for 5 min at 80°C each sample was divided into two separate series: one for HpaII, one for. The DNA samples were digested with 1 µl of FastDigest^® HpaII (Fermentas) and 1 µl of FastDigest^® MspI (Fermentas) in a total volume of 50 µl for 30 min at 37°C. Enzymes were inactivated by incubation for 15 min at 65°C. EcoRI and HpaII/MspI adapters (Table 1) were prepared by mixing 50 µM oligonucleotides in equal volumes and incubation at the following profile: 10 min at 65°C, 10 min at 37°C, 10 min at 25°C, 10 min at 4°C. The restricted DNA was added the ligation mixture (2 µM Eco and Hpa/Msp adapter, 1×ligase buffer, 0.2 mM dATP, 5U T₄ DNA ligase (Fermentas)) and incubated for 3 h at 37°C. The samples were diluted ten times, and 5 µl of the diluted DNA was added to the preamplification mixture. The total volume of 50 µl of preamplification mixture also contained: 10 µM Hpa + A/C primer (Table 1), 10 µM Eco + A primer (Table 1), 1×polymerase buffer, 0.2 mM dNTP, 2.5 mM MgCl₂, 2 U DNA Polymerase (DyNAzyme™, Finnzymes, Espoo, Finland). The PCR conditions were as follows: initial denaturation of 1 min at 94°C; 25 cycles—30 s at 94°C, 60 s at 60°C, 60 s at 72°C; final elongation of 5 min at 72 °C. The PCR products were diluted 20 times and 5 µl of the diluted sample was added to 15 µl of selective amplification mixture. The final concentrations of the components in the mixture were: 4 µM Hpa selective primer (Table 1), 4 µM Eco selective primer (Table 1), 1 × polymerase buffer, 0.2 mM dNTP, 2.5 mM MgCl₂, 0.5 U DNA Polymerase (DyNAzyme™, Finnzymes). The PCR conditions were as follows: initial denaturation of 1 min at 94°C; 12 cycles—30 s at 94°C, 30 s at 60°C, 60 s at 72°C; 22 cycles—30 s at 94°C, 30 s at 56 °C, 60 s at 72°C final elongation of 30 min at 72°C. In total 35 different combinations of selective primers were tested, out of which nine combinations were chosen for the final analysis (Table 1).

PCR products were separated on a 6.5% denaturing polyacrylamide gel. 100 ml of the gel was prepared from RapidGel™-XL-6% Liquid Acrylamide (Affymetrix, Cleveland, USA) containing 6% modified acrylamides, 7 M Urea, 89 mM Tris, 89 mM boric acid and 2 mM EDTA, and to this solution 1,000 µl of 10% APS and 100 µl of TEMED was added. Electrophoresis in 1×TBE buffer was carried out on the Sequi-Gen GT Sequencing Cell (Bio-Rad, Hercules, USA) with the Power Pac™(Bio-Rad) as a power supply unit. The gel was pre-electrophoresed for 1 h (100 W, 2,000 V, 50 mA). Loading buffer (98% formamide, 20 mM EDTA, 0,01% bromophenol blue, 0,01% xylene cyanol) was added to the samples (12 µl per 20 µl), the mixture was denatured at 95 °C for 5 min, placed in ice for 10 min. 7 µl of the mixture from each samples were loaded in the wells. Simultaneously with the samples the pBR322 DNA/BsuRI (HaeIII) marker (Fermentas) was prepared and loaded onto gel. Electrophoresis was carried out for 2.5 h (100 W, 2,000 V, 50 mA). After electrophoresis bands were detected by silver staining technique using Silver Sequence™ DNA Sequencing System (Promega, Madison, USA), according to manufacturer’s protocol. The documentation and analysis of the gels were done by means of GS-800 Calibrated Densitometer (Bio-Rad) and Quantity One^® ver. 4.6.9 (Bio-Rad) software.

MSAP procedure was repeated—three biological replicates were performed. Reproducible, clearly distinguishable fragments were scored in a form of binary matrix as either present (1) or absent (0) for the samples. MSAP markers were classified as different types: 11, 01, 10, or 00 accordingly to the presence (1) or absence (0) of the corresponding amplified fragment in the digestion reactions with HpaII and MspI, respectively.

Assessment of methylation level (%) of Methylation-Susceptible Loci (MSL), Principal Coordinate Analyses (PCoA) and Neighbor-Joining tree were done by means of R program (R version 3.2.2) with msap package.

UPGMA (unweighted pair group method using arithmetic averages algorithm was used for cluster analysis in order to show the relationships between taxa. Dice Index was computed in FreeTree (Pavlícek, Hrdá & Flegr, 1999) program and the dendrogram was visualized in TreeView (Page, 1996) program. Bootstrapping with 1,000 repetition was done in Free Tree program. The data from double matrix including: 11, 01, 10 and 00 types were transformed into single 1/0 matrix, in which 1 represented any type of methylated fragment (types: 01, 10 or 00) and 0 represented unmethylated fragment (type 11).

The analysis of differences between the taxa (U-values) was based on the suggestion of Wang et al. (2014): ${\displaystyle p=\frac{y1+y2}{n1+n2};}$ ${\displaystyle q=1-p;}$ ${\displaystyle {\delta }_{p1-p2}=\surd \left(pq\left(1∕n1\right)+1∕n2\right);}$ ${\displaystyle U=\frac{p1-p2}{{\delta }_{p1-p2}}}$

where n1 represents the total number of fragments of a given sample, n2 the total fragments of another sample, y1 the total number of methylated fragments of a given sample values, y2 the total number of methylated fragments in another sample, p1 the proportion (%) of methylated fragments of a given sample values and p2 the proportion (%) of methylated fragments in another sample.

Giemsa staining

Roots from 2-day-old seedlings were treated with 0.05% colchicine for 3 h at room temperature in the dark prior to fixation in ethanol-glacial acetic acid (3:1). Roots were macerated in a mixture of 4% (w/v) pectinase (Fluka), 6% (w/v) hemicellulase (Sigma-Aldrich, St. Louis, USA) and 4% (w/v) cellulase (Sigma-Aldrich) in 0.01 M citric acid—sodium citrate buffer (pH 4.8), for 3 h at 37°C . Roots were washed in distilled water. Each root tip were dissected on a single microscope slide, a drop of 45% acetic acid was added and the root tip was squashed under a cover glass. Preparations were heated for 20 min at 47 °C. The cover slip was removed after freezing over dry ice, and the slides were air-dried overnight. Chromosomes were stained with Giemsa reagent according to Darvey & Gustafson (1975) and Merker (1976). Preparations were analyzed under Eclipse E600 microscope (Nikon, Tokyo, Japan). NIS Elements ver. 3.00 SP7 (Nikon) software was used to measure: total chromosomes lengths, arms lengths and telomeric C-bands lengths. Based on the measurements the percentage of total telomeric heterochromatin content in the genome of each Secale taxon was computed. The Kruskal–Wallis test was performed in Statistica software ver. 12.

ELISA

The 5-mc DNA ELISA Kit (Zymo Research, Irvine, CA, USA) was used to detect global 5-methylocytosine in the DNA. 100 ng of each sample DNA was used in the assay. Four replicates of each sample were prepared in a single experiment, and the experiment was repeated three times. The analysis was performed according to the manufacturer’s instruction. Adequate controls included in the were used in the experiments. Plates were read in the Infinite^® M200 Pro microplate reader (Tecan, Crailsheim, Germany). Magellan™ ver. 7.0 (Tecan) software was used to generate the standard curve and to quantify the percentage of 5-mC in DNA samples. The Tuckey test was performed in Statistica software ver. 12.

Flow cytometry—20 mg of coleoptiles from 7-day-old seedlings were chopped in 1 ml of ice-cold nuclei isolation buffer (45 mM MgCl₂⋅6H₂O, 20 mM MOPS, 0,1% (v/v) Triton X-100, 1% PVP; pH 7.0) supplemented with propidium iodide (Sigma-Aldrich; 50 µg/mL) and ribonuclease A (Sigma-Aldrich; 50 µg/mL). The homogenate was filtered through a 42-mm nylon mesh into a 1.5 ml tube and incubated for 10 min at ice. Secale cereale was used as an internal standard. The measurements were done at the rate 20–50 nuclei. For each sample, fluorescence in at least 7,000 was measured using BD FACSCalibur™ flow cytometer (Becton Dickinson, San Jose, CA, USA). The mass of the genome was calculated from the formula below: ${\displaystyle \text{Sample}2C\text{value}\left(pg\right)=\text{Reference}2C\text{value}\times \frac{\text{Sample}2C\text{mean}\text{peak}\text{position}}{\text{Reference}2C\text{mean}\text{peak}\text{position}}.}$ Reference 2C value for Secale cereale = 16.55 pg (Bennett & Smith, 1976).

Histograms were analyzed using the CellQuest™ Pro (Becton Dickinson) software. The results were estimated using Tuckey test (Statistica software ver. 12).

MSAP

Methylation profiling by MSAP relies on the use of methylation-sensitive enzymes. In this work we used a pair of enzymes, MspI and HpaII, which recognize the 5^′-CCGG-3′ sequence, but they cut the DNA depending on the methylation state of that sequence: MspI will not cut the DNA if the external cytosine is fully methylated or hemimethylated. HpaII will not cut the DNA when the cytosine is fully methylated, and it cuts the DNA when the external cytosine is hemimethylated (Table 2). Therefore, the application of this pair of enzymes in MSAP allows to specify 4 different methylation states of the 5′-CCGG-3′ sequence: type 11—no methylation, type 01—methylated CpG sequence, type 10—hemimethylated sequence or methylated CpCpG sequence, type 00—hypermethylation (Table 2). In type 4, it is also possible that no band in a given taxon does not result from changes in the methylation state, the reason may be the lack of this sequence. Hence, the MSAP analysis may introduce some error. However, based on the conducted experiments, including less clear bands omitted in this analysis, it could be said that band patterns of the examined rye taxa were very similar to each other, and even if the analysis was burden with the above-described error, it was small. Before we performed the MSAP analysis, some preliminary tests with the classical AFLP technique had been done (A Kalinka, 2014, unpublished data). We calculated the Dice coefficient for each pair of species/subspecies. The values were high (0.89–0.95), thus we decided that MSAP analysis would be reliable. The values of the coefficient were not surprising as from ISSR and IRAP analyses, which allow to study the polymorphism of highly variable inter-microsatellite and inter-retrotransposon sequences, the values were not low, 0.48–0.85 (IRAP) and 0.39–0.82 (ISSR) (Achrem, Kalinka & Rogalska, 2014). We were also aware that MSAP can analyze a small quantity of methylated cytosines in the genome that is why we also complemented with global methylation data, which shed more light on the variation of methylation in Secale.

To ensure the robustness of the results, the analysis included only the most evident and repetitive fragments in the size range 50–500 nt. In total, 149 loci were selected of all the reactions performed with 9 primer combinations. Among these loci, 98 methylation-susceptible loci (MSL) and 51 non-methylated loci (NML) were distinguished. The number of polymorphic MSL was 14 (14% of total MSL) and the NML was 7 (14% of total NML). The level of methylation of MSL loci is shown in Fig. 1, which demonstrates that type 01 (internal methylated cytosine) and type 00 (hypermethylation) fragments are most frequent among rye taxa, reaching approximately 43% and 38%, respectively. Type 11 fragments were definitely the least frequent (approximately 7%), and were followed by slightly more frequent type 10—(approx. 12%).

Both the MSL analysis alone (Fig. 2) and the combined analysis of MSL and NML loci (Fig. 3) showed that S. sylvestre was the most distinct from other species. Secale vavilovii also showed greater epigenetic variation in relation to the taxa belonging to the species S. cereale and S. strictum. The total methylation level of the studied loci was very similar in all taxa (Fig. 4). The U values were calculated between all taxa to verify whether the differences in the level of methylation are statistically significant. Higher U values indicated greater differences between taxa, but only U values >1.96 (U_(0.05) = 1.96) would mean statistically significant differences. The U values in the studied rye taxa ranged between 0 for S. sylvestre and S. vavilovii, S. sylvestre and S. cereale, S. cereale and S. vavilovii and 1.25 for Secale s. ssp. africanum and S.c. ssp. segetale, indicating that the methylation pattern did not differ significantly between any pair of taxa (Table 3). Dice coefficients confirmed this high similarity (Table 3), as it ranged between pairs of individual taxa from 87% (S. sylvestre and S. s. ssp. kuprijanovii) to 96% (S. cereale and S. s. ssp. kuprijanovii, S. cereale and S.c. ssp. segetale, S.c. ssp. segetale and S. s. ssp. kuprijanovii). Since the rye taxa investigated in this study belonged to 4 species, it was very interesting how epigenetically similar to each other were subspecies within the species. Such analysis was conducted and the graphical representation is shown in Fig. 2. Figure 2A shows the variation between the taxa without grouping them in species, while Fig. 2B shows the variation between species. Dendrogram (Fig. 2C) shows the relationship between eight rye taxa grouped into four species. The results clearly indicated that epigenetic similarity between species was sufficiently significant so that the subspecies of S. cereale and S. strictum did not group together. The bootstrapping analysis confirmed this result (Fig. 3). Low values obtained indicated that the presented clustering of taxa was unreliable, and only S. sylvestre was definitely different from other taxa, but these differences were not statistically significant.

Giemsa staining

Thirty preparations of root tips of each rye taxa were prepared in order to determine the total percentage of telomeric heterochromatin. Five preparations were analyzed in detail, in which nine best metaphase plates were selected and the following measurements were performed in them: the overall length of each chromosome, short and long arm length, the length t-heterochromatin bands in individual chromosomes. On this basis, the total percentage of telomeric heterochromatin was calculated in the whole genome of each taxon. The results showed that the content of t-heterochromatin was varied (Fig. 4): the lowest amount of t-heterochromatin was found in S. sylvestre (6.18%) (Fig. 5B), S. s. ssp. africanum (6.25%), S. strictum (7.18%) and S.s. ssp. kuprijanovii (7.72%). No significant differences in the content of t-heterochromatin were found between these taxa (Table 4). More t-heterochromatin was present in the genomes of S. c. ssp. segetale (10.94%), S. c. ssp. afghanicum (11.90%), S. cereale (12.29%) and S. vavilovii (13.39%) (Fig. 5A). Also, in this case no significant differences were recorded between these taxa. Therefore, these Secale taxa could be divided into two groups, with a relatively lower and higher content of t-heterochromatin. In contrast to the MSAP method, the results of which showed that there are no significant differences between all taxa, and the taxa of the species S. cereale and S. strictum did not cluster together, the analysis of the content of t-heterochromatin confirmed the currently accepted taxonomic relationships. Admittedly, species that were the most different from others in the MSAP analysis, here also had the lowest (S. sylvestre; 6.18%) and the highest (S. vavilovii; 13.39%) t-heterochromatin contents. While the remaining taxa in the group with smaller amounts of heterochromatin-t belonged to the species S. strictum, and in the group with a higher t-heterochromatin content they belonged to the species S. cereale. Nevertheless, there were no statistically significant differences found in the t-heterochromatin content between S. s. ssp. kuprijanovii and S. c. ssp. segetale and S. c. ssp. afghanicum.

ELISA

In total, 12 measurements of the global cytosine content were performed in the genome of each of the rye taxa. Five taxa had a similar level of methylated cytosine (Fig. 4) and no significant differences were observed (Table 4) between S.c. ssp. afghanicum (74.41%), S. s. ssp. africanum (69.52%), S. cereale (69.80%), S. s. ssp. kuprijanovii (67.05%) and S. vavilovii (66.42%). Slightly lower level of methylation was detected in S. sylvestre (61.09%) and it only differed from the methylation state of S. c. ssp. afghanicum of the above-listed taxa. Very high content of 5-methylcytosine (82.86%) was found in S. c. ssp. segetale. In this case, it has a significantly higher level of methylation compared to the genomes of all the other taxa, with the exception of S. c. ssp. afghanicum. On the other hand, extremely low levels of DNA methylation was present in S. strictum (52.56%) and differed significantly from the methylation status of all taxa besides S. sylvestre.

The level of global cytosine methylation in DNA, as opposed to the total t-heterochromatin in the genome, did not overlap with the current taxonomic relationships. Within three taxa belonging to the species S. cereale: S. cereale, S. c. ssp. afghanicum and S. cereale differed significantly from S. c. ssp. segetale, while in the taxa belongong to S. strictum: subspecies S. strictum subsp. strictum were different from the other two subspecies S. s. ssp. africanum and S. s. ssp. kuprijanovii.

It is not surprising that the results of methylation analysis of the CCGG sequence by MSAP did not overlap with global analysis of cytosine methylation by ELISA, as in plant genomes, in addition to the methylation of CG sequences, CNG and CNN sequence methylation also occurs (Johnson et al., 2007; Meyer, 2011; González, Ricardi & Iusem, 2013). Thus, it was possible that the CpG methylation was similar in the genomes of all Secale taxa tested, while such high differences in the level of 5-methylcytosine were due to differences in the CNG and CNN methylation. In addition, the MSAP method allowed to examine only selected loci in a limited number, which might not give a complete picture of epigenetic variation.

Flow cytometry

The DNA content in the nucleus of the studied Secale taxa was the lowest in S. s. ssp. africanum (14.76 pg) and S. s. ssp. kuprijanovii (15.23 pg), in which it did not exceed 16 pg. All other taxa had the 2C value above 16 pg, with the largest amount of DNA measured in S. vavilovii (Fig. 6). It was this rye species that differed significantly from S. s. ssp. africanum (P < 0.001) and S. s. ssp. kuprijanovii (P = 0.007). In other cases, there were no significant differences in the DNA content in the nucleus (P > 0.05).