Skip to main content

Analysis of the leaf methylomes of parents and their hybrids provides new insight into hybrid vigor in Populus deltoides

Abstract

Background

Plants with heterosis/hybrid vigor perform better than their parents in many traits. However, the biological mechanisms underlying heterosis remain unclear. To investigate the significance of DNA methylation to heterosis, a comprehensive analysis of whole-genome DNA methylome profiles of Populus deltoides cl.'55/65' and '10/17' parental lines and their intraspecific F1 hybrids lines was performed using methylated DNA immunoprecipitation (MeDIP) and high-throughput sequencing.

Results

Here, a total of 486.27 million reads were mapped to the reference genome of Populus trichocarpa, with an average unique mapping rate of 57.8%. The parents with similar genetic background had distinct DNA methylation levels. F1 hybrids with hybrid vigor possessed non-additive DNA methylation level (their levels were higher than mid-parent values). The DNA methylation levels in promoter and repetitive sequences and transposable element of better-parent F1 hybrids and parents and lower-parent F1 hybrids were different. Compared with the maternal parent, better-parent F1 hybrids had fewer hypermethylated genes and more hypomethylated ones. Compared with the paternal parent and lower-parent L1, better-parent F1 hybrids had more hypermethylated genes and fewer hypomethylated ones. The differentially methylated genes between better-parent F1 hybrids, the parents and lower-parent F1 hybrids were enriched in the categories metabolic processes, response to stress, binding, and catalytic activity, development, and involved in hormone biosynthesis, signaling pathway.

Conclusions

The methylation patterns of the parents both partially and dynamically passed onto their hybrids, and F1 hybrids has a non-additive mathylation level. A multidimensional process is involved in the formation of heterosis.

Background

Heterosis/hybrid vigor is the phenomenon in which progeny are superior to their parents (with distinct genetic backgrounds) in many traits, such as biomass, growth rate, adaptability, fertility, and resistance [1–5]. Since interspecific hybrid tobacco with hybrid vigor was produced in the 1760s by crossing Nicotiana rustuca with N. paniculata [6], heterosis has often been exploited in crop and tree breeding. However, the genetic basis of heterosis is still far from being understood and is still a controversial subject [1, 7, 8]. Three classic hypotheses, i.e., dominance, overdominance, and epistasis, were proposed as genetic explanations for heterosis. In the dominance hypothesis, the inferior parental alleles in the hybrids are complemented by the superior or dominant alleles from the other parent. In the overdominance hypothesis, heterosis arises from allelic interactions within each of many genetic loci. An alternate model, epistasis, postulates that interactions between different parental genes in hybrids lead to heterosis. Although numerous examples support each of these hypotheses, they only partially explain the genetic basis of heterosis [1, 3, 9].

Genome-wide analyses of heterosis have revealed altered gene expression profiles in F1 hybrids compared with their parents, as well as non-additive patterns of gene expression [2, 10], including studies in maize (Zea mays L.) [11–14], rice (Oryza.sativa L.) [15, 16], Arabidopsis thaliana [17], wheat (Triticum aestivum L.) [18], Larix kaempferi (Lamb.) Carr [19], and Populus tremula [20]. Recently, heterosis was observed in hybrids derived from parents with similar genetic backgrounds. Such parents with highly similar genomic features had distinct epigenomes [21–23], and epi-alleles that arise from epigenetic modification were also identified. Epi-alleles cause allelic variation and altered gene expression activity, which are essential to the architecture of plant heterosis [24]. One type of epigenetic regulation, DNA methylation, primarily serves as an epigenetic silencing mechanism and predominantly occurs in transposons and other repetitive DNA elements [5, 25–29] and has been explored in model plants and crops, such as maize, rice, cotton (Gossypium herbaceum L.), and A. thaliana.

The genus Populus (poplar) includes species that are important for the health of ecosystems and are vital to the timber, paper, and biofuel industries. Poplars are also used as a model woody plant species and models of interest for epigenetic studies [30, 31]. Variations in DNA methylation between genotypes and tissues and in response to drought, as well as the relationship between gene-body DNA methylation and tissue-specific gene expression, have been reported [31–34].

During the last century, many poplar varieties with enhanced growth or adaptability have been generated using inter- or intraspecific hybridization approaches, which take advantage of the presence of heterosis in poplars. Although investigations of the molecular basis of heterosis in poplar have been undertaken via genetic mapping and gene expression profiling, the global patterns of epigenetic modification such as DNA methylation have not been determined, and whether DNA methylation plays a role in the architecture of heterosis is still unclear. In this study, P. deltoides cl.'55/65' was maternal parent which has straight bole, round crown, fast growth, high resistance to Anoplophora glabripennis and strong rooting ability, and P. deltoides cl. '10/17' was paternal parent which fast-growth and high stress resistance. This cross-combination is multigeneration convergent cross. Intraspecific F1 hybrids of P. deltoides with significant hybrid vigor or lower-parental performance were examined. Methylated DNA immunoprecipitation, combined with a high-throughput sequencing (MeDIP-Seq) approach were applied to analyze the genome-wide DNA methylation landscapes in Populus deltoides parental lines and F1 hybrids lines. The results showed that better-parent F1 hybrids have higher methylation levels than the average of the parents, suggesting that non-additive level of DNA methylation is related to heterosis/hybrid vigor. The hypermethylated genes of better-parent F1 hybrids relative to the parents and lower-parent F1 hybrids were enriched in the processes of metabolism and development, which may be highly relevant to heterosis.

Methods

Plant materials and growth conditions

Two P. deltoides intraspecific parental lines, P. deltoides cl. '55/65' (Salicaceae, Populus, Section Aigeiros) and P. deltoides cl. '10/17' (Salicaceae, Populus, Section Aigeiros) and their intraspecific hybrids, designated here as H1, H2, H3, L1 and L2, were used in this study. All F1 Hybrids was generated by the same intraspecific cross-combination of P. deltoides cl. '55/65' as maternal parent and P. deltoides cl. '10/17' as paternal parent. P. deltoides cl. '55/65' was primitively generated from the inbred seeds of excellent individual plants in former Yugoslavia and introduced into China in 1981. P. deltoides cl. '10/17' was generated by intraspecific crossing P. deltoides Bartr. cv. 'Shanhaiguanensis' (which was primitively generated from the inbred seeds of excellent individual plants and introduced into China in 1900) with P. deltoides Bartr. cl. 'Harvard' (I-63/51) (which was primitively generated from the inbred seeds of excellent individual plants in Mississippi Delta and introduced into China in 1972).

Hybrids were generated by hand pollination. All seeds were grown in a greenhouse at the Chinese Academy of Forestry (the authority responsible is the Chinese Academy of Forestry, Beijing, China) in January, 2002. One-year-old seedlings were made into cuttings to accelerate cloning, which were planted in the greenhouse in January, 2003 and transplanted to Yuquan mountain nursery (the authority responsible is the Chinese Academy of Forestry, Beijing, China) in May, 2003. No specific permits were required for these locations. The locations are not privately owned in any way, and the field studies did not involve endangered or protected species. A total of 149 F1 hybrids were introduced into Jiaozuo Research Institute of Forestry (Henan province, China) in 2003 and 2004. Of these, 18 F1 hybrids that had good performance in tree height and Diameter at breast height (DBH) were selected over the course of the two-year seedling test. Parents and their 18 F1 hybrids were planted in Xifeng village, Wuzhi Country, Jiaozuo city in Henan province in 2005 and then transplanted to Yangcheng, Wuzhi Country, Jiaozuo city of Henan province (35°8' N, 113°17' E), in 2007. No specific permits were required for these locations. The location is not privately owned in any way, and the field studies did not involve endangered or protected species. This site has an annual average precipitation of 625.4 mm, with an annual average temperature of 15.2ºC (ranging from 14.3ºC to 43.6ºC), an accumulated temperature above 0ºC of 4,633ºC, and a frostless period of 224 days per year. The average relative humidity and annual sunshine duration are 61% and 2,434 hours, respectively. The experimental field had an average soil pH of 6.8 and was irrigated. This trial was designed in randomized complete blocks, with four blocks and eight trees per treatment (planting spacing of 3 m × 5 m). After 5 years of growth, three F1 hybrids (H1, H2, and H3) which exhibited the highest tree heights and largest DBHs and two F1 hybrids (L1 and L2) that showed the lowest tree heights and DBHs were selected.

Since DNA methylation differences among tissues are obvious in Poplar [34] and leaves are important to plant growth and development, after five years of growth, the leaves at the top of main trunk were collected at the vigorous stage (9:30-10:30 am on August 10, 2011). Three trees (three leaves per tree) per replication were sampled, thus, twelve trees and 36 leaves were sampled for every line. Samples for every parent and F1 hybrid were pooled and stored in liquid nitrogen prior to DNA extraction.

Evaluation of heterosis

Since planting (in 2007), two important economic traits, tree height and DBH were continuously measured. Considering heterosis over higher parent was important for poplar breeding, after five years of growth, heterosis over higher parent was calculated as H = (F1-Ps)/Ps × 100%, where H is the amount of heterosis, F1 is the trait value measured in the hybrid, and Ps is the trait value measured in the higher parent [35].

MeDIP-Seq

Genomic DNA was isolated from each sample using a DNeasy Plant Mini Kit (Qiagen, Courtaboeuf, France). The DNA integrity was verified by agarose gel electrophoresis. The DNA was quantified using a Qubit Fluorometer and a Quant-iTâ„¢ dsDNA BRAssay Kit (Life Technologies, USA).

The MeDIP process was almost identical to the method of Pomraning et al [36]. Before carrying out MeDIP, genomic DNA was sheared to 350-450 bp fragments with a Bioruptor (Sonics, Newtown, USA, VC130PB), and the fragments were recovered using a Qiaquick PCR Purification Kit (Qiagen, Courtaboeuf, France). The fragments were end-repaired, phosphorylated, and A-tailed. The fragments were then ligated to Illumina sequencing adapters [37]. The sheared DNA was diluted in 450 µl of TE buffer, denatured in a 100°C heat block for 10 min, and snap-cooled on ice for 5 min. Immunoprecipitation buffer (100 mM Na-Phosphate pH 7.0, 1.4 M NaCl, 0.5% TritonX-100) and 1 µl of 5meC antibody (Diagenode, Liège, Belgium #MAb-5MECYT-100, 1 µg/µl) were added to the DNA solution followed by incubation for 2 h on an orbital rotator at 4°C. Bound DNA was precipitated with sheep anti-mouse IgG Dynabeads (M-280, Invitrogen, California,USA), washed three times with immunoprecipitation buffer for 10 min at room temperature with shaking, resuspended in 250 μl proteinase K digestion buffer (5 mM Tris, pH 8.0, 1 mM EDTA, pH 8.0, 0.05% SDS) with 7 μl of 10 mg/ml proteinase K, and incubated for 3 h on an end-over-end rotator at 50°C to digest the antibodies and release the 5meC-containing DNA. Methylated DNA was extracted by phenol-chloroform extraction followed by ethanol precipitation. The DNA pellets were resuspended in 50 μl TE buffer and stored at -20°C.

The immunoprecipitated DNA was used to generate a DNA colony template library using the Fasteris procedure (Fasteris, Plan-les-Ouates, Switzerland). The DNA samples were quantified using a 2100 Bioanalyzer (Agilent, USA) and a StepOnePlus Real-Time PCR System (ABI, California,USA). Illumina sequencing was performed in a HiSeq-2000 system (Illumina, San Diego, CA, USA).

Bioinformatics processing and statistical analysis

MeDIP-Seq reads were aligned to the Populus trichocarpa v2.2 reference genome (http://www.phytozome.net/poplar.php, February 2012). The alignments were carried out with SOAP aligner (BGI, version 2.01) [38], allowing up to two mismatches for successful mapping. The mapped rate (the ratio of the number of mapped reads to that of original reads), and the uniquely mapped rate (the ratio of the number of uniquely mapped reads to that of original reads) were calculated. The coverage depth was calculated as the coverage times of specific loci by sequencing reads. The genome coverage was calculated as the proportion of eligible base numbers in the entire genome. In the distribution analysis of the MeDIP-Seq sequencing reads in a chromosome, each chromosome was scanned with windows of 100 kb, the reads coverage depth per window was calculated, and the reads were standardized with the following formula: reads number of specific 100 kb windows * 1,000,000/number of uniquely mapped reads. The methylation coverage of CG/CHG/CHH contexts was calculated as the proportion of CG/CHG/CHH site over certain coverage depth in all CG/CHG/CHH sites from as determined by MeDIP-Seq.

Peak summit coordinates were generated using model-based analysis of ChIP-Seq (MACS; version 1.4.0 beta) [39]. The summit files were then used for further analysis (total peaks number, peak mean length, peak median length, peak total length, and peak covered size in the genome).

To detect differentially methylated gene between the two samples, the peak summits of two samples were merged, and the normalized reads number of each sample the merged region was determined. The false positive reads were removed using a chi-square test. For genes that overlapped with a merged region, if the reads number of sample 2 in this region was more than that of sample 1, then the gene was designated as hypermethylated during the Sample 1 versus Sample 2 comparison, while if the opposite situation occurred, the gene was considered to be hypomethylated.

Gene Ontology (GO) analysis was performed to obtain the functional classifications of differentially methylated genes using the TermFinder tool (http://search.cpan.org/~sherlock/GO-TermFinder-0.86/). P-values were multiple test corrected to reduce false positive rates. GO terms with adjusted P-values of <0.05 were considered to be significant.

The known genes were submitted to the KEGG Automatic Annotation Server (http://www.genome.jp/kegg/pathway.html) for pathway analysis. A hypergeometric test was performed to identify the significantly enriched pathways in differentially methylated genes compared with the whole genome. Pathways with Q-values ≤ 0.05 were considered to be significant.

Results

Heterosis performance

Analysis of Variance (ANOVA) of the height and DBH of trees of various ages (from one to five years old) for the parents and F1 hybrids was performed. The results showed that the tree height and DBH in each age of parents and F1 hybrids were significantly different (Figure 1, Additional file 1: Table S1). The tree heights and DBHs of H1, H2, and H3 at each age were higher than those of the parents (except for the annual tree height of H2). The tree heights of H1, H2, and H3 at five-year were significantly greater than those of the parents. The tree heights of L1 and L2 at five-year were lower than those of the parents, and the DBH of L1 and L2 at five-year were both significantly lower than those of the parents. Since heterosis over higher parent is important for poplar breeding, we estimated the heterosis over higher parent values (Table 1). Hybrids H1, H2, and H3 exhibited heterosis over higher parent for tree height (7.81% for H1, 12.55% for H2, and 11.09% for H3) and DBH (1.26% for H1, 1.49% for H2, and 0.72% for H3), while hybrids L1 and L2 possessed negative heterosis over higher parent for tree height (-5.77% for L1 and -7.59% for L2) and DBH (-20.92% for L1 and -21.82% for L2).

Figure 1
figure 1

Growth traits of the parents and their F1 hybrids. Tree height (A) and DBH (B) were measured during five successive years. Means are given with ± SE. Different lower case letters indicate significant differences (P < 0.05) among the lines in trees of the same age. Identical letter indicate the absence of a significant difference and different letter indicate significant difference.

Table 1 Heterosis of F1 hybrids

Mapping of MeDIP-Seq reads to the reference genome

MeDIP-Seq libraries were constructed from DNA extracted from the parents and their hybrids and subjected to high-throughput Solexa sequencing (Illumina, San Diego, CA, USA). A total of 670 million reads were produced from the P1, P2, H1, H2, H3, L1, and L2 lines. The reads were mapped to the P. trichocarpa v2.2 reference genome (http://www.phytozome.net/poplar.php). A total of 486.27 million reads could be mapped in this process. The mapped rates for the seven lines were 83.4%, 74.2%, 82.4%, 65.1%, 65.9%, 63.5%, and 71.7%, respectively (average of 72.31%). The uniquely mapped rates were 66.4%, 56.8%, 65.6%, 55.3%, 52.0%, 53.0%, and 55.6%, respectively (average of 57.8%; Table 2). To facilitate the access and use of the P. deltoides methylome sequencing data, the raw data in the FASTQ format was deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database with accession number SRP034728.

Table 2 Summary of MeDIP-Seq experimental results

Comparison of methylation status among parents and F1 hybrid genomes

In this study, the leaf methylomes of the parents and F1 hybrids were investigated. The distribution of MeDIP-Seq reads on the 19 scaffolds (each scaffold represents a putative chromosome) of Populus was shown in Figure 2. Distinct DNA methylomes were observed among the parents and F1 hybrids. Among the 19 chromosomes, nine (I, IV, VI, VII, XI, XII, XVI, XVIII, and XIX) had greater methylation coverage in the middle parts of chromosomes, which may comprise the centromeric regions [34]. Moreover, three chromosomes (II, VIII, and XIII) had greater methylation coverage in distal parts of the chromosomes, while four (III, IX, X, and XV) had greater methylation coverage in proximal parts of the chromosomes.

Figure 2
figure 2

Chromosome-level view of methylation in parents and F1 hybrids. MeDIP-Seq reads were plotted in 100-kb windows along each chromosome. One line is shown for each line.

Different types of methylation coverage were also observed in the parents and F1 hybrids (Figure 3). In plant, DNA methylation in often found in all sequence, namely, the symmetric CG and CHG contexts (H represents A, T, and C) and asymmetric CHH contexts [40]. We calculated the methylation coverage of CG/CHG/CHH contexts. We found that remarkable methylation occurred in each cytosine context in the genomes of parents and F1 hybrids. Moreover, the methylation coverage in CG, CHH, and CHG displayed clear differences between parents P1 and P2. The methylation coverage in the three sequence contexts for maternal parent P1 was 17.24%, 19.24%, and 18.83%, respectively, while that for paternal parent P2 was 12.61%, 13.35%, and 12.92%, respectively (Figure 3). For better-parent hybrids, the methylation coverage of H1 was 16.97% (CG), 18.79% (CHG), and 18.44% (CHH), respectively. H2 and H3 had comparable values to H1, with an average of 16.06%, 17.40%, and 17.33% for CG, CHG, and CHH contexts, respectively (Figure 3). These results indicate that the methylation levels of all better-parent hybrids were between those of the two parents (less than P1 and more than P2), while the values were higher than the average levels of the parents (14.94%, 16.31%, and 15.88%, respectively). For lower-parent hybrids, L2 displayed an average methylation coverage of 16.35% for the three contexts, which was less than that of P1 but more than that of P2, while the values of L1 were 10.21%, 10.21%, and 10.05%, respectively, which were less than those of both parents (Figure 3).

Figure 3
figure 3

Comparison of DNA methylationin F1 hybrids relative to their parents. Each column indicates the proportion of the number of CG/CHG/CHH context over 1X sequence depth to all number of CG/CHG/CHH contexts. Black line indicates thepercentage of CG contexts in better parent, and black dotted line indicates the average of percentag of CG contexts in two parents.

Mapping of MeDIP-Seq reads to genes

The distribution of MeDIP-Seq reads on various genomic features, including CpG islands, promoters, 5' untranslated regions (UTRs), 3' UTRs, coding sequences (CDS), and introns, was characterized based on methylation coverage. Promoters are defined as the 2-kb region upstream of the annotated transcription start site. Methylation coverage in promoters of the parents and F1 hybrid was higher than that of the gene body (Figure 4). Moreover, CDS and introns of the gene body had higher methylation coverage, while the 5'-and 3'- UTRs had very low methylation coverage. We also compared the methylation coverage on various genomic features among parents, better-parent F1 hybrids, and lower-parent F1 hybrids. We found that three better-parent F1 hybrids had higher methylation coverage in promoters, 5 'UTRs and 3' UTRs than those of the parents and the lower-parent hybrids. In intron, the methylation coverage of H2 was higher than other lines, the coverage of P1, H1, H3 and L2 were similar, the coverage of P2 and L1 were similar (Figure 4, Additional file 2: Figure S1). H2 had higher methylation coverage in CDS, whereas the coverage of L2 was higher than H1 and H3 and the parents.

Figure 4
figure 4

Fraction of methylation of genome features between parents and hybrids. Each column indicates the proportion of reads distributed in specific genome features of the total mapped reads. Promoters are defined as the 2-kb region upstream of the annotated transcription start site. CGI was determained by CpGPlot (http://www.ebi.ac.uk/Tools/seqstats/emboss_cpgplot/help/index.html).

CpG islands (CGI) are a prime target for epigenetic modification. Moreover, CpG islands are more frequently found in plant genomes than in the human genome [41]. CGIs were previously thought to be unmethylated unless they were located at genomic imprinting sites or on the inactivated × chromosome [26]. Recent studies have shown that some CpG islands are methylated [42, 43]. To further observe the distribution trend of DNA methylation in CGIs of Populus, 2,000-bp regions upstream and downstream of CGI were divided into 20 segments, and the CGI was divided into 40 segments. By counting the normalized average coverage depth, we determined that the CGIs in the parents and hybrids had methylation. H2 had the highest level of CGI methylation, followed by the lower-parent hybrids L1 and L2, the paternal parent P2, H1 and H3, and the maternal parent P1 (Figure 5).

Figure 5
figure 5

The trend of average coverage depth of DNA methylation in CGI. The region within the dotted line indicates CGI region.

Transposable elements and repetitive sequences are frequently methylated regions. Previous reports have suggested that inferring the methylation status in transposons and repetitive sequences at the whole-genome level using classic approaches, especially microarray- and PCR-based assays, is unreliable. This problem can be overcome by using the recently available MeDIP-Seq approach, which can be used to deduce the coverage of all major types of methylation for transposons and repetitive sequences [44]. In plants, numerous LTR-gypsy retrotransposon elements are present in the heterochromatic centromeric and pericentromeric regions. LTR-gypsy retrotransposon elements are the most abundant type of transposon element in the Populus trichocarpa genome [45]. We detected the enrichment of LTR-gypsy retrotransposons in the DNA-methylated fraction of the genomes of the parents and F1 hybrids. In addition, genome regions containing LTR copia, DNA/En-Spm, Low-complexity (which contains a highly non-uniform amino acid composition [46, 47]), and Simple-repeat were also methylated (Figure 6). The variations in methylation in transposable elements and repetitive sequences seem dependent on each genotype. The methylation coverages of the F1 hybrids were between those of the two parents in LTR-Gypsy, simple-repeat and Low_complexity (except for L2 in LTR-Gypsy). In LTR-Copia, H1, H3 and L2 had higher methylation coverage with H2 and L1 had lower methylation coverage. In contrary, in rRNA and DNA/En-Spm, H2 and L1 had higher methylation coverage, whereas the methylation coverage of H1 and H3 were lower.

Figure 6
figure 6

Fraction of methylation of repetitive sequence and transposable element between parents and hybrids. Each column indicates the proportion of reads distributed in a specific element of the total mapped reads. Information of repeat element was obtained from the RepeatMasker (http://www.repeatmasker.org/faq.html).

Methylated peaks analysis

To avoid the false positive results generated from MeDIP-Seq, model-based analysis of ChIP-Seq (MACS) was used to obtain methylation peak summits. In this analysis, the total number of peaks of the better-parent hybrids was higher than that of the parents, while that of the lower-parent hybrids was lower than that of the parents. The parents had distinct DNA methylation peak coverage. The number of peak coverage of the better-parent hybrid H1 was more than that of the parents, and those of the other two better-parent hybrids (H2 and H3) were higher than the average levels of the parents. Moreover, the number of peak coverages of both lower-parent hybrids was lower than those of the low parents (Table 3).

Table 3 Statistics of peak summits

We further mapped peaks to various genomic features. Table 4 shows the peak number contained in each genomic feature. The peak number in the promoter regions was greater than that of the gene body. In the gene body, CDS had a higher peak number than introns and UTRs. The peak numbers in the promoter, 5'UTR, and 3'UTR of better-parent hybrid F1 plants were higher than those of the parents. In introns, the peak number of L1 was higher than that of the other lines, while the peak numbers of P1, H1, and H2 were similar, and the peak numbers of P2, H3, and L2 were also similar.

Table 4 Number of peaks in genome features

Analysis of differentially methylated genes in the parental and F1 hybrid genomes

The number of differentially methylated gene among the parents and F1 hybrids is shown in Figure 7. Better-parent hybrid H1 had a similar number of hyper- and hypomethylated genes as the maternal parent P1 (589 and 580, respectively). By contrast, remarkably fewer genes with hypermethylated were detected relative to hypomethylated genes in better-parent hybrid H2 (1,338 versus 4,189 genes) and H3 (414 versus 1,046 genes). Compared with the paternal parent P2, the better-parent hybrids displayed higher levels of hypermethylation at protein-coding genes, as more hypermethylated than hypomethylated genes were found in H1 (2,887 versus 1,054 genes), H2 (1,854 versus 1,496 genes), and H3 (1,734 versus 1,668 genes). However, there were less hypermethylated than hypomethylated genes in hybrid L1 (681 versus 2,181 genes). When better-parent and lower-parent F1 hybrids were compared, we found higher levels of methylation in the better-parent hybrids than in L1, as inferred by the larger number of hypermethylated genes than hypomethylated genes in H1 (5,551 versus 1,113 genes), H2 (2,479 versus 514 genes), and H3 (4,002 versus 1,041 genes). Compared with L2, more genes with hypermethylation than hypomethylated genes were detected in H1 (2,985 versus 1,253 genes) and H2 (1,196 versus 1,071 genes), whereas H3 showed higher levels of methylation, with a greater number of hypomethylated genes (1,135 hypermethylated versus 1,330 hypomethylated).

Figure 7
figure 7

Differentially methylated genes between parents and hybrids. Light gray represents hypomethylated genes, and dark gray represents hypermethylated genes. The selective conditions were, p < 0.01, two samples had coverage in the same genome feature, and the coverage difference was more than two times.

For all of the differentially methylated genes identified, we performed Gene Ontology (GO) functional category analysis to determine whether these genes were enriched for certain pathway or network (Additional file 2: Figure S2). The results showed that the differentially methylated genes between better-parent hybrid H1 and maternal parent P1 were enriched in 28 biological functional categories, and ten additional enriched functional categories (biological adhesion, cell proliferation, locomotion, reproductive process, extracellular region, extracellular region part, enzyme regulator activity, molecular transducer activity, protein binding, and transcription factor activity) were also found for genes identified in the H1-P2 comparison. Compared with maternal parent P1, better-parent hybrid H2 possessed more hypermethylated genes enriched in 35 functional categories, such as biological adhesion, cell proliferation, protein binding, and transcription factor activity. The differentially methylated genes between H2 and paternal parent P2 were enriched in 33 functional categories (e.g., pigmentation). The differentially methylated genes between better-parent hybrid H3 and maternal parent P1 were enriched in 31 functional categories, and two additional categories (cell proliferation and molecular transducer activity) were found to have enriched differentially methylated genes between H3 and P2. As a whole, the majority of hypermethylated genes between three better-parent hybrids and both parents tended to fall into seven functional categories, including metabolic processes, cellular, response to stress, cell, cell part, binding, and catalytic activity.

To further investigate the differentially methylated genes between better-parent F1 hybrids and the parents, we analyzed hypermethylated genes enriched in specific functional categories in H1, H2, and H3 versus the two parents (Figure 8A). 20 and 199 genes showed hypermethylation in all three better-parent F1 hybrids compared with P1 and P2, respectively. Among these genes, 97 fell into seven major GO functional categories, namely metabolic process (31), cellular metabolic process (21), primary metabolic process (17), small molecule metabolic process (10), nitrogen compound metabolic process (7), developmental process (6), and anatomical structure development (5). Notably, four genes (POPTR_0008s18650, POPTR_0010s02290, POPTR_0017s03190, and POPTR_0010s19920) with hypermethylation were detected in all better-parent hybrid-parent comparisons (Table 5). Among these, POPTR_0010s02290 encodes a predicted GTP-binding protein with GTPase activity and protein binding functions, and POPTR_0010s19920 encodes 3-dehydrosphinganine reductase, which participates in metabolic processes and oxido-reductase activity. The genes were submitted to the KEGG Pathway (http://www.genome.jp/kegg/pathway.html) database to obtain their KEGG orthology annotations. POPTR_0010s19920 was found to participate in sphingolipid metabolism (ko00600). No annotation was retrieved for POPTR_0008s18650 or POPTR_0017s03190, suggesting that these genes have unknown functions.

Figure 8
figure 8

Comparison of differentially methylated genes in various genome features among better-parent hybrids, the parents, and lower-parent hybrids. (A) The number of differentially methylated genes between better-parent hybrids and the maternal parents were calculated (dark gray columns), and the number of differentially methylated genes between better-parent hybrids and the paternal parents are indicated in the light gray columns. (B) The number of differentially methylated genes between better-parent hybrids and below-parent hybrids L1 were calculated (dark gray columns), and the number of differentially methylated genes between better-parent hybrids and below-parent hybrids L2 are indicated in the light gray columns.

Table 5 Annotations of hypermethylated genes

The differentially methylated genes between better- and lower-parent F1 hybrids were also analyzed (Figure 8B). The number of hypermethylated genes of better-parent F1 hybrids versus lower-parent hybrids L1 and L2 were 523 and 132, respectively. For these genes, hypermethylation predominantly occurred in the promoter (171 genes compared with L1 and 27 genes compared with L2) and gene body (296 genes compared with L1 and 80 genes compared with L2), while less hypermethylation occurred in the 5 'UTR (28 genes compared with L1 and 15 genes compared with L2) and the 3' UTR (15 genes compared with L1 and 10 genes compared with L2). A total of 40 hypermethylated genes were found in all three better-parent F1 hybrids compared with lower-parent hybrids L1 and L2, ten of these genes could be annotated (Table 5). The GO functional categories of the 10 genes mainly involve metabolic process (4), primary metabolic process (2), cellular metabolic process (3), signaling (3), small molecule metabolic process (2), anatomical structure development (2), and biological regulation (2). These genes were then submitted to the KEGG Pathway database, yielding pathway information about four genes as follows: (1) POPTR_0012s07360 is calcium-dependent protein kinase gene involved in plant-pathogen interactions (ko04626); (2) POPTR_0015s09720 belongs to cytochrome P450 CYP4/CYP19/CYP26 subfamilies involved in steroid hormone biosynthesis (ko00140); (3) POPTR_0019s09910 encodes a molybdopterin synthase catalytic subunit involved in multiple processes such as metabolism, metabolism of cofactors, vitamin and folate biosynthesis (ko00790), genetic information processing, folding, sorting, and degradation, and the sulfur relay system (ko04122); and (4) POPTR_0008s18420 encodes an erbb2-interacting protein involved in a NOD-like receptor signaling pathway (ko04621).

Discussion

Several classical hypotheses about heterosis are based on the differences between genomes [1], and allelic diversity may produce heterosis. However, hybrid vigor can be observed even when parents are genetically very similar [24]. Recent studies have shown that parents with similar genome sequences have distinct epigenomes, which may contribute to heterosis [5, 24]. In Populus, hybrids with heterosis are often obtained by intrasection and interspecific hybridization, whereas hybrids obtained by intersection hybridization always have mid-parent performance, and hybrids with growth vigor are obtained less frequently. In Section Aigeiros, excellent hybrids with heterosis have been produced by intraspecific hybridization; the level of heterosis increases with a decrease in genetic distance between parents and polymerization of excellent genetic composition. Super high yield varieties are often generated by convergent crossing of P. deltoides varieties (strains). In this study, P. deltoides cl. '55/65', was used as the maternal parent and P. deltoides cl. '10/17' was used as the paternal parent. This cross combination is a multigeneration convergent cross, and the level of heterosis is outstanding. Intraspecific F1 hybrids of P. deltoides with significant hybrid vigor or lower-parental performance were examined, providing a unique opportunity to accurately analyze the contribution of DNA methylation to heterosis in trees. This is the first investigation of DNA methylation maps with high resolution in P. deltoides plants and their F1 hybrids at the genome-wide scale using high-throughput sequencing.

A total of 670.55 million reads were generated using MeDIP-Seq, 486.27 million of which could be mapped onto the reference genome of P. trichocarpa, the average of the uniquely mapped rates was 57.8%. The relatively low rate of mapping using genomes of closely related species as a reference suggests that species in different sections within the genus Populus are genetically divergent (P. trichocarpa belongs to sect. Tacamahaca, and P. deltoides belongs to sect. Algeiros). Similar observations were also documented in studies of Populus alba and Populus tremula based on single nucleotide polymorphism (SNP) analysis of the two species [48]. Our dataset of leaf methylomes shows that the parents and F1 hybrids had significant methylation in the CG/CHH/CHG contexts, with CHG and CHH methylation being more consistent, and cytosines in CG context were less methylated than those in the other two contexts. Previous studies have reported that CGs are dominant in methylome, especially in coding regions, while less frequent in general, 5meCHH is more common in repeat regions and short transposable elements [49–51]. In populous, CG and CHG methylation were more consistent within tissues. However, in the two targets with cytosine content < 10%, cytosines in CHH context were methylated more frequently than those in the other two contexts [34]. The two parents had distinct methylomes reflected by different methylation coverage in the CG/CHG/CHH contexts. The methylation coverage of three better-parent F1 hybrids was higher than the average of the parental values (mid-parent value, MPV), indicating that the F1 hybrids had an altered epigenome, and the DNA methylation level was non-additive. Unlike in animal systems, where "Erase and Reset" of cytosine methylation occurs in each generation, in plants, the parental methylation states can be stably inherited by the progeny [52, 53]. However, many plants species often exhibit the remodeling of parental methylation patterns in interspecific hybrids and allopolyploids [54–56]. In these scenarios, DNA methylation partly functions epigenetically and dynamically over generations, thus achieving the control and balance of gene expression under specific circumstances [27, 54, 57].

Early studies proposed that allelic variation is the primary cause of heterosis [58], but this notion was challenged by the observation that parents with similar genetic backgrounds can also produce hybrids with heterosis, which can arise from the diversity of epialleles. Epi-allelic changes in hybrids occur though changes in siRNA levels, trans-chromosomal methylation (TCM) or trans-chromosomal demethylation (TCdM), which fit the dominance or overdominance hypotheses and indicate that epi-alleles are essential parts of the genetic basis of heterosis. In rice hybrids, DNA methylation at many loci is inherited by non-additive inheritance [29]. Although the two rice hybrids had unequal numbers of non-additively methylated loci, in both hybrids, approximately 75% of such loci had increased methylation levels. The increased DNA methylationwas also reported in reciprocal F1 hybrids between Arabidopsis thaliana Landsberg erecta and C24 [5]. In this study, we found that P. deltoides F1 hybrids with hybrid vigor (H1, H2, and H3) showed higher DNA methylation coverage in three contexts than the MPV. This can partially be explained by the effects of TCM. In this scenario, the better parent derived siRNA molecules associate with both alleles, maintains the methylation state of its own alleles and establishes the de novo methylation of lower parent hypomethylation [24], resulting in increased methylation levels in the non- or low methylation region. Therefore, the methylation levels of hybrids may exceed MPV. For lower-parent hybrids L1, the fact that DNA methylation coverages in three contexts are lower than parental values can be attributed to the influence of TCdM. The lower parent derived siRNA initially becomes associated across both parental alleles. This association can cause siRNA level to be present at lower levels than the threshold required for the establishment and/or maintenance of methylation, leading to hypomethylation of alleles of the lower parent allele. At the same time, with the loss of methylation, normal siRNA levels cannot be maintained (loss of siRNA), which further reduces the level of DNA methylation, as detected in the lower-parent hybrids L1 with lower methylation levels. Thus, contrasting patterns of methylation between poplar better-parent F1 and lower-parent L1 hybrids may result from an adjustment of methylation levels of the parents, and this difference in methylation may in turn influence and regulate the expression network of target genes, which is beneficial to the establishment of heterosis. Interestingly, one of the hybrids with negative better-parent heterosis (L2) has methylation coverages in three contexts above the midparent value, and the variations in methylation in specific genomic features (such as intron) and in transposable elements and repetitive sequences seem dependent on each genotype. This indicates that the role of DNA methylation in heterosis is complex and multifaceted.

In addition, in some annual herb plant species, distinct epigenomes between parents can give rise to increased DNA methylation levels in the F1 hybrids and contribute to heterosis. For instance, when two rice subspecies, Nipponbare (o. sativa ssp japonica) and 93-11 (o. sativa ssp indica), were used as parents, 82.1 and 70.8% of the different methylation region (DMRs) of the genome of F1 hybrids showed high- or above high-parental DNA methylation levels, respectively [29]. When A. thaliana Landsberg erecta and C24 were used as parental lines, the reciprocal F1 hybrids showed increased DNA methylation levels across the entire genome, especially in the transposable elements [5].

However, other studies revealed no obviously altered or decreased methylation levels in hybrids compared with their parents. In Arabidopsis thaliana, 97% of the MspI/HpaII recognition sites in the F1 hybrids of a Col-0 and C24 cross retained their levels of methylation [59]. The methylation levels of cotton hybrids were lower than those of the parents, and the demethylation numbers of better-parent hybrids were higher than those of the lower-parent hybrids [60]. This discrepancy may be due to the different approaches used in these two studies versus the present study. The two previous studies used a methylation-sensitive amplified polymorphism assay, which is much less sensitive than MeDIP-Seq and thus could not fully scan all methylation loci and could only partially provide the landscapes of DNA methylation.

The MACS approach can improve the spatial resolution of the aligned data and impart the robustness of the final aligned sequences based on dynamic Poisson distribution [39]. The peak coverage further illustrates that the parents had distinct DNA methylation levels, while F1 hybrids with hybrid vigor possessed elevated DNA methylation levels, and F1 hybrids with negative hybrid vigor possessed declining DNA methylation levels. In the P. deltoides genomes, peak data were found to be more enriched in promoters than in gene bodies, and the CDS showed more enrichment than introns or UTRs in gene bodies. The enrichment levels of various genomic features in the better-parent hybrids, parent and lower-parent hybrid were different. The growth vigor displayed in better-parent hybrids may be attributed to the increased transcriptional inactivation of CG and CHG sites and heterochromatin-mediated gene silencing, which are related to methylated enrichment. Throughout the growth and development of poplar, methylated enrichment may also suppress the expression of a proportion of genes and/or reduce spurious global transcription to enable full transcription or to initiate the expression of other suitable loci, consequently increasing hybrid vigor in the F1 hybrids; this concept deserves further investigation.

The analysis of differentially methylated genes between parents and hybrids has revealed that the hypermethylation levels of better-parent F1 hybrids were between those of the two parents, while the hypermethylation levels of the lower-parent F1 hybrids was lower than lowest value of the parents. This finding suggests that having a methylation level between that of the two parents in F1 hybrids may be more favorable for achieving better-parent heterosis, while deviating from the MPV tends to preclude the establishment of heterosis.

The analysis of GO functional categories showed that the differentially methylated genes between the better-parent F1 hybrids and the parents were enriched in metabolic processes, response to stress, and binding and catalytic activity, which indicates that heterosis in trees may follow a comprehensive process. At the same time, compared with lower-parent F1 hybrids, the hypermethylated genes in the better-parent F1 hybrids were enriched in metabolic and development processes, such as metabolic process, cellular metabolic process, primary metabolic process, small molecule metabolic process, nitrogen compound metabolic process, developmental process, anatomical structure development, and signaling, which implied that differentially methylated genes are involved in heterosis.

Compared with the parents and lower-parent F1 hybrids, the hypermethylated genes in better-parent F1 hybrids were involved in hormone synthesis and response to stress, such as cytochrome P450, participating in the biosynthesis of hormones, defensive compounds and fatty acids, GTP-binding proteins involved in cytoskeleton organization, signal transduction, vesicle trafficking, and stress tolerance. As Ca2 + signal transducers, calcium-dependent protein kinases play an important role in various plant physiological process, including growth, development, defense responses, regulation of reactive oxygen species production, symbiotic interactions, guard cell turgor, osmotic, drought and salt stress, and regulation through hormones such as ABA and GA. In summary, the fact that many differentially methylated genes are involved in diverse biological pathways indicates that the inheritance of heterosis is a multidimensional process.

Conclusions

To date, studies linking epigenetics and heterosis have only been carried out in a few plant species. In this study, we identified genome-wide variations in leaf methylomes between parents and their hybrids in P. deltoides, a perennial forest tree species. The dataset derived from MeDIP-Seq were used to produce DNA methylation maps with high resolution of P. deltoides. cl. '55/65' and P. deltoides cl. '10/17' and their five F1 hybrids. Populus F1 hybrids has a non-additive mathylation level (higher than mid-parent values), which showed that the methylation patterns of the parents partially and dynamically passed onto their hybrids and was remodeled. In addition, the DNA methylomes of better-parent F1 hybrids were significantly different from that of lower-parent F1 hybrids, which indicates that having a methylation level between that of the two parents may be more favorable for the achievement of better-parent heterosis in F1 hybrids, while the deviation from MPV tends to preclude the establishment of heterosis. Compared with the parents and the lower-parent F1 hybrids, the hypermethylated genes in the better-parent F1 hybrids were enriched in the processes of metabolism and development, which may be highly relevant to heterosis.

References

  1. Birchler JA, Auger DL, Riddle NC: In search of the molecular basis of heterosis. Plant Cell. 2003, 15: 2236-2239. 10.1105/tpc.151030.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  2. Birchler JA, Yao H, Chudalayandi S, Vaiman D, Veitia RA: Heterosis. Plant Cell. 2010, 2105-2112. 22

  3. Hochholdinger F, Hoecker N: Towards the molecular basis of heterosis. Trends in Plant Science. 2007, 427-432. 12

  4. Chen ZJ: Molecular mechanisms of polyploidy and hybrid vigor. Trends in Plant Science. 2010, 15: 57-71.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  5. Shen H, He H, Li J, Chen W, Wang X, Guo L, Peng Z, He G, Zhong S, Qi Y, et al: Genome-wide analysis of DNA methylation and gene expression changes in two arabidopsis ecotypes and their reciprocal hybrids. Plant cell. 2012, 24: 875-892. 10.1105/tpc.111.094870.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  6. Roberts: Plant Hybridization before Mendel. 1929, Princeton, NJ: Princeton University Press

    Book  Google Scholar 

  7. Reif JC, Warburton ML, Xia XC, Hoisington DA, Crossa J, Taba S, Muminović J, Bohn M, Frisch M, Melchinger AE: Grouping of accessions of Mexican races of maize revisited with SSR markers. Theoretical and Applied Genetics. 2006, 177-185. 113

  8. Nakamura S, Hosaka K: DNA methylation in diploid inbred lines of potatoes and its possible role in the regulation of heterosis. Theoretical and Applied Genetics. 2010, 205-214. 120

  9. He G, Elling AA, Deng XW: The epigenome and plant development. Annual Review of Plant Biology. 2010, 62: 411-435.

    Article  Google Scholar 

  10. Lippman ZB, Zamir D: Heterosis: revisiting the magic. Trends in Genetics. 2007, 23: 60-66. 10.1016/j.tig.2006.12.006.

    Article  CAS  PubMed  Google Scholar 

  11. Swanson-Wagner RA, Jia Y, DeCook R, Borsuk LA, Nettleton D, Schnable PS: All possible modes of gene action are observed in a global comparison of gene expression in a maize F1 hybrid and its inbred parents. Proceedings of the National Academy of Sciences, USA. 2006, 6805-6810. 103

  12. Swanson-Wagner RA, DeCook R, Jia Y, Bancroft T, Ji T, Zhao X, Nettleton D, Schnable PS: Paternal dominance of trans-eQTL influences gene expression patterns in maize hybrids. Science. 2009, 1118-1120. 326

  13. Paschold A, Marcon C, Hoecker N, Hochholdinger F: Molecular dissection of heterosis manifestation during early maize root development. Theoretical and Applied Genetics. 120: 441-450.

  14. Meyer S, Pospisil H, Scholten S: Heterosis associated gene expression in maize embryos 6 days after fertilization exhibits additive, dominant and overdominant pattern. Plant Molecular Biology. 2007, 63: 381-391. 10.1007/s11103-006-9095-x.

    Article  CAS  PubMed  Google Scholar 

  15. Zhang HY, He H, Chen LB, Li L, Liang MZ, Wang XF, Liu XG, He GH, Chen RS, Ma LG, et al: A genome-wide transcription analysis reveals a close correlation of promoter INDEL polymorphism and heterotic gene expression in rice hybrids. Molecular Plant. 2008, 1: 720-731. 10.1093/mp/ssn022.

    Article  CAS  PubMed  Google Scholar 

  16. Wei G, Tao Y, Liu G, Chen C, Luo R, Xia H, Gan Q, Zeng H, Lu Z, Han Y, et al: A transcriptomic analysis of superhybrid rice LYP9 and its parents. Proceedings of the National Academy of Sciences, USA. 2009, 106: 7695-7701. 10.1073/pnas.0902340106.

    Article  CAS  Google Scholar 

  17. Andorf S, Selbig J, Altmann T, Poos K, Witucka-Wall H: Enriched partial correlations in genome-wide gene expression profiles of hybrids (A. thaliana): A systems biological approach towards the molecular basis of heterosis. Theoretical and Applied Genetics. 2010, 120: 249-259. 10.1007/s00122-009-1214-z.

    Article  CAS  PubMed  Google Scholar 

  18. Wang Z, Ni Z, Wu H, Nie X, Sun Q: Heterosis in root development and differential gene expression between hybrids and their parental inbreds in wheat (Triticum aestivum L.). Theoretical and Applied Genetics. 2006, 113: 1283-1294. 10.1007/s00122-006-0382-3.

    Article  CAS  PubMed  Google Scholar 

  19. Li A, Fang MD, Song WQ, Chen CB, Qi LW, Wang CG: Gene expression profiles of two intraspecific Larix lines and their reciprocal hybrids. Molecular Biology Reports. 2012, 4: 3773-3784.

    Article  Google Scholar 

  20. Zhuang Y, Adams KL: Extensive allelic variation in gene expression in Populus F1 hybrids. Genetics. 2007, 177: 1987-1996. 10.1534/genetics.107.080325.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  21. Jin H, Hu W, Wei Z, Wan L, Li G, Tan G, Zhu L, He G: Alterations in cytosine methylation and species-specific transcription induced by interspecific hybridization between Oryza sativa and O. officinalis. Theoretical and Applied Genetics. 2008, 117: 1271-1279. 10.1007/s00122-008-0861-9.

    Article  CAS  PubMed  Google Scholar 

  22. Takamiya T, Hosobuchi S, Noguchi T, Asai K, Nakamura E, Habu Y, Paterson AH, Iijima H, Murakami Y, Okuizumi H: Inheritance and alteration of genome methylation in F1 hybrid rice. Electrophoresis. 2008, 29: 4088-4095. 10.1002/elps.200700784.

    Article  CAS  PubMed  Google Scholar 

  23. Zhang MS, Yan HY, Zhao N, Lin XY, Pang JS: Endosperm-specific hypomethylation, and meiotic inheritance and variation of DNA methylation level and pattern in sorghum (Sorghum bicolor L.) inter-strain hybrids. Theoretical and Applied Genetics. 2007, 115: 195-207. 10.1007/s00122-007-0555-8.

    Article  CAS  PubMed  Google Scholar 

  24. Groszmann M, Greaves IK, Albertyn ZI, Scofield GN, Peacock WJ, Dennis ES: Changes in 24-nt siRNA levels in Arabidopsis hybrids suggest an epigenetic contribution to hybrid vigor. Proceedings of the National Academy of Sciences, USA. 2011, 108: 2617-2622. 10.1073/pnas.1019217108.

    Article  CAS  Google Scholar 

  25. and Colot V: DNA methylation and epigenetic inheritance in plants and filamentous fungi. Science. 2001, 1070-1074. 293

  26. Bird A: DNA methylation patterns and epigenetic memory. Genes & Development. 2002, 6-21. 16

  27. Chan SW, Henderson IR, Jacobsen SE: Gardening the genome:DNA methylation in Arabidopsis thaliana. Nature Reviews Genetics. 2005, 6: 351-360.

    Article  CAS  PubMed  Google Scholar 

  28. Law JA, Jacobsen SE: Establishing, maintaining and modifying DNA methylation patterns in plants and animals. Nature Reviews Genetics. 2010, 204-220. 11

  29. He G, Zhu X, Elling A, Chen L, Wang X, Guo L, Liang M, He H, Zhang H, Chen F, et al: Global epigenetic and transcriptional trends among two rice subspecies and their reciprocal hybrids. Plant Cell. 2010, 22: 17-33. 10.1105/tpc.109.072041.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  30. Hamanishi ET, Campbell MM: Genome-wide responses to drought in forest trees. Forestry. 2011, 273-283. 84

  31. Lafon-Placette C, Faivre-Rampant P, Delaunay A, Street N, Brignolas F, Maury S: Methylome of DNase I sensitive chromatin in Populus trichocarpa shoot apical meristematic cells: a simplified approach revealing characteristics of gene-body DNA methylation in open chromatin state. New Phytologist. 2013, 197: 416-30. 10.1111/nph.12026.

    Article  CAS  PubMed  Google Scholar 

  32. Gourcilleau D, Bogeat-Triboulot M-B, Le Thiec D, Lafon-Placette C, Delaunay A, El-Soud WA, Brignolas F, Maury S: DNA methylation and histone acetylation: genotypic variations in hybrid poplars, impact of water deficit and relationships with productivity. Annals of Forest Science. 2010, 67: 208-10.1051/forest/2009101.

    Article  Google Scholar 

  33. Raj S, Brautigam K, Hamanishi ET, Wilkins O, Thomas BR, Schroeder W, Mansfield SD, Plant AL, Campbell MM: Clone history shapes Populus drought responses. Proceedings of the National Academy of Sciences, USA. 2011, 108: 12521-12526. 10.1073/pnas.1103341108.

    Article  CAS  Google Scholar 

  34. Vining KJ, Pomraning KR, Wilhelm LJ, Priest HD, Pellegrini M, Mockler TC, Freitag M, Strauss SH: Dynamic DNA cytosine methylation in the Populus trichocarpa genome: tissue-level variation and relationship to gene expression. BMC Genomics. 2012, 13: 27-10.1186/1471-2164-13-27.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  35. Xiao J, Li J, Yuan L, McCouch SR, Tanksley SD: Genetic diversity and its relationship to hybrid performance and heterosis in rice as revealed by PCR-based markers. Theoretical and Applied Genetics. 1996, 92: 637-643. 10.1007/BF00226083.

    Article  CAS  PubMed  Google Scholar 

  36. Pomraning KR, Smith KM, Freitag M: Genome-wide high throughput analysis of DNA methylation in eukaryotes. Methods. 2009, 47: 142-150. 10.1016/j.ymeth.2008.09.022.

    Article  CAS  PubMed  Google Scholar 

  37. Sati S, Tanwar VS, Kumar KA, Patowary A, Jain V, Ghosh S, Ahmad S, Singh M, Chandak GR, et al: High Resolution Methylome Map of Rat Indicates Role of Intragenic DNA Methylation in Identification of Coding Region. PLoS ONE. 2012, 7 (2): e31621-10.1371/journal.pone.0031621. doi:10.1371/journal.pone.0031621

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  38. Li R, Yu C, Li Y, Lam TW, Yiu SM, Kristiansen K, Wang J: SOAP2: an improved ultrafast tool for short read alignment. Bioinformatics. 2009, 25: 1966-1967. 10.1093/bioinformatics/btp336.

    Article  CAS  PubMed  Google Scholar 

  39. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nussbaum C, Myers RM, Brown M, Li W, et al: Model-based analysis of ChIP-Seq (MACS). Genome Biology. 2008, 9: R137-10.1186/gb-2008-9-9-r137.

    Article  PubMed Central  PubMed  Google Scholar 

  40. Henderson IR, Jacobsen SE: Epigenetic inheritance in plants. Nature. 2007, 418-424. 447

  41. Ashikawa I: The Gene-associated CpG islands in plants as revealed by analyses of genomic sequences. The Plant Journal. 2001, 26: 617-625. 10.1046/j.1365-313x.2001.01062.x.

    Article  CAS  PubMed  Google Scholar 

  42. Yamada Y, Watanabe H, Miura F, Soejima H, Uchiyama M, Iwasaka T, Mukai T, Sakaki Y, Ito T: A comprehensive analysis of allelic methylation status of CpG islands on human chromosome 21q. Genome Research. 2004, 247-266. 14

  43. Straussman R, Nejman D, Roberts D, Steinfeld I, Blum B, Benvenisty N, Simon I, Yakhini Z, Cedar H: Developmental programming of CpG island methylation profiles in the human genome. Nature Structural & Molecular Biology. 2009, 564-571. 16

  44. Feber A, Wilson GA, Zhang L, Presneau N, Idowu B, Down TA, Rakyan VK, Noon LA, Lloyd AC, Stupka E, et al: Comparative methylome analysis of benign and malignant peripheral nerve sheath tumors. Genome Research. 2011, 21: 515-524. 10.1101/gr.109678.110.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  45. Douglas C, DiFazio S: Genetics and Genomics of Populus. In The Populus genome and comparative genomics. Edited by: Jansson S, Bhalerao R, Groover A. 2010, Berlin: Springer-Verlag, 67-90.

    Chapter  Google Scholar 

  46. Wootton JC, Federhen S: Statistics of local complexity in amino acid sequences and sequence databases. Computers & chemistry. 1993, 17 (2): 149-163. 10.1016/0097-8485(93)85006-X.

    Article  CAS  Google Scholar 

  47. Price AL, Jones NC, Pevzner PA: De novo identification of repeat families in large genomes. Bioinformatics. 2005, 21 (suppl 1): i351-i358. 10.1093/bioinformatics/bti1018.

    Article  CAS  PubMed  Google Scholar 

  48. Stölting KN, Nipper R, Lindtke D, Caseys C, Waeber S: Genomic scan for single nucleotide polymorphisms reveals patterns of divergence and gene flow between ecologically divergent species. Molecular Ecology. 2012, 22: 842-855.

    Article  PubMed  Google Scholar 

  49. Cokus SJ, Feng S, Zhang X, Chen Z, Merriman B, Haudenschild CD, Pradhan S, Nelson SF, Pellegrini M, Jacobsen SE: Shotgun bisulphite sequencing of the Arabidopsis genome reveals DNA methylation patterning. Nature. 2008, 452 (7184): 215-219. 10.1038/nature06745.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  50. Widman N, Jacobsen SE, Pellegrini M: Determining the conservation of DNA methylation in Arabidopsis. Epigenetics. 2009, 4 (2): 119-124. 10.4161/epi.4.2.8214.

    Article  CAS  PubMed  Google Scholar 

  51. Cokus SJ, Feng S, Zhang X, Chen Z, Merriman B, Haudenschild CD, Pradhan S, Nelson SF, Pellegrini M, Jacobsen SE: Shotgun bisulphite sequencing of the Arabidopsis genome reveals DNA methylation patterning. Nature. 2008, 452 (7184): 215-219. 10.1038/nature06745.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  52. Cubas P, Vincent C, Coen E: An epigenetic mutation responsible for natural variation in floral symmetry. Nature. 1999, 401: 157-161. 10.1038/43657.

    Article  CAS  PubMed  Google Scholar 

  53. Kakutani T: Epi-alleles in plants: inheritance of epigenetic information over generations. Plant and Cell Physiology. 2002, 43: 1106-1111. 10.1093/pcp/pcf131.

    Article  CAS  PubMed  Google Scholar 

  54. Madlung A, Masuelli RW, Watson B, Reynolds SH, Davison J, Comai L: Remodeling of DNA methylation and phenotypic and transcriptional changes in synthetic Arabidopsis allotetraploids. Plant Physiology. 2002, 129: 733-746. 10.1104/pp.003095.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  55. Lukens LN, Pires JC, Leon E, Vogelzang R, Oslach L, Osborn T: Patterns of sequence loss and cytosine methylation within a population of newly resynthesized Brassica napus allopolyploids. Plant Physiology. 2006, 140: 336-348.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  56. Marfil CF, Masuelli RW, Davison J, Comai L: Genomic instability in Solanum tuberosum Solanum kurtzianum interspecific hybrids. Genome. 2006, 49: 104-113.

    CAS  PubMed  Google Scholar 

  57. Zhang MS, Yan HY, Zhao N, Lin XY, Pang JS: Endosperm-specific hypomethylation and meiotic inheritance and variation of DNA methylation level and pattern in sorghum (Sorghum bicolor L.) inter-strain hybrids. Theoretical and Applied Genetics. 2007, 115: 195-207. 10.1007/s00122-007-0555-8.

    Article  CAS  PubMed  Google Scholar 

  58. Sakthivel K, Girishkumar K, Ramkumar G, Shenoy VV, Kajjidoni ST: Alterations in inheritance pattern and level of cytosine DNA methylation, and their relationship with heterosis in rice. Euphytica. 2010, 175: 303-314. 10.1007/s10681-010-0167-2.

    Article  CAS  Google Scholar 

  59. Banaei MAM, Fuchs J, Czauderna T, Houben A, Mette MF: Intraspecific hybrids of Arabidopsis thaliana revealed no gross alterations in endopolyploidy, DNA methylation, histone modifications and transcript levels. Theoretical and Applied Genetics. 2010, 120: 215-226. 10.1007/s00122-009-1127-x.

    Article  Google Scholar 

  60. Zhao Y, Yu S, Xing C, Fan S, Song M: DNA methylation in cotton hybrids and their parents. Mlekuliarnaia biologiia. 2008, 195-205. 42

Download references

Declarations

The publication charges this article were funded by the National Key Technology Research and Development Program for the Twelfth Five-Year-Plan of China (Grant No. 2012BAD01B03) and National Key Technology Research and Development Program for the Eleventh Five-Year-Plan of China (Grant No. 2006BAD01A15).

This article has been published as part of BMC Genetics Volume 15 Supplement 1, 2014: Selected articles from the International Symposium on Quantitative Genetics and Genomics of Woody Plants. The full contents of the supplement are available online at http://www.biomedcentral.com/bmcgenet/supplements/15/S1.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Xiaohua Su.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

GM and HQ carried out the experiments and performed manuscript draft writing. GM, HQ and CY performed bioinformatics and statistical analysis. DC and ZB participated in the design and coordination the study. SX designed the study and revised this manuscript. All authors read and approved the final manuscript.

Ming Gao, Qinjun Huang contributed equally to this work.

Electronic supplementary material

12863_2014_1258_MOESM1_ESM.doc

Additional file 1: Additional file 1 includes Table S1, which gives detailed information about the growth comparisons between the parents and F1 hybrids. Tree height and DBH were measured during five successive years. Means are given with ± SE. Different letters indicate significant difference (P < 0.05) among the lines in trees of the same age. (DOC 40 KB)

12863_2014_1258_MOESM2_ESM.pdf

Additional file 2: Additional file 2 includes Figure S1 and Figure S2. Figure S1 describes the trend of average coverage depth of DNA methylation in the intragenic region. The region within the dotted line indicates intragenic region. 2,000-bp regions upstream and downstream of intragenic are divided into 20 segments, and the intragenic are divided into 40 segments. Figure S2 provides details of GO analysis of differentially methylated genes. (A) GO analysis of differentially methylated genes between the maternal parent and better-parent F1 hybrids. Three comparison pairs (H1 versus P1, H2 versus P1, and H3 versus P1 were included in this analysis. (B) GO analysis of differentially methylated genes between the paternal parent and better-parent F1 hybrids. H1 versus P2, H2 versus P2, and H3 versus P2 were included in this analysis. (C) GO analysis of differentially methylated genes between the maternal parent and lower-parent F1 hybrids. Two comparison pairs (L1 versus P1, L2 versus P1) were included in this analysis. (D) GO analysis of differentially methylated genes between the paternal parent and lower-parent F1 hybrids. L1 versus P2 and L2 versus P2 were included in this analysis. (PDF 853 KB)

Rights and permissions

Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.

The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/.

The Creative Commons Public Domain Dedication waiver (https://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Gao, M., Huang, Q., Chu, Y. et al. Analysis of the leaf methylomes of parents and their hybrids provides new insight into hybrid vigor in Populus deltoides. BMC Genet 15 (Suppl 1), S8 (2014). https://doi.org/10.1186/1471-2156-15-S1-S8

Download citation

  • Published:

  • DOI: https://doi.org/10.1186/1471-2156-15-S1-S8

Keywords