Horticulture Research, 2022, 9: uhac127 https://doi.org/10.1093/hr/uhac127 Article Absence of major epigenetic and transcriptomic changes accompanying an interspecific cross between peach and almond Carlos de Tomás1 , Amélie Bardil2 , Raúl Castanera1 , Josep M. Casacuberta1 , * and Carlos M. Vicient1 , * 1 Centre for Research in Agricultural Genomics CSIC-IRTA-UAB-UB, Campus UAB, Edifici CRAG, Bellaterra, Barcelona 08193, Spain 2 Present address: Institut écologie et environnement (INEE), CNRS, Montpelier, France *Corresponding authors. E-mail: josep.casacuberta@cragenomica.es; carlos.vicient@cragenomica.es Abstract Hybridization has been widely used in breeding of cultivated species showing low genetic variability, such as peach (Prunus persica). The merging of two different genomes in a hybrid often triggers a so-called “genomic shock” with changes in DNA methylation and in the induction of transposable element expression and mobilization. Here, we analysed the DNA methylation and transcription levels of transposable elements and genes in leaves of Prunus persica and Prunus dulcis and in an F1 hybrid using high-throughput sequencing technologies. Contrary to the “genomic shock” expectations, we found that the overall levels of DNA methylation in the transposable elements in the hybrid are not significantly altered compared with those of the parental genomes. We also observed that the levels of transcription of the transposable elements in the hybrid are in most cases intermediate as compared with that of the parental species and we have not detected cases of higher transcription in the hybrid. We also found that the proportion of genes whose expression is altered in the hybrid compared with the parental species is low. The expression of genes potentially involved in the regulation of the activity of the transposable elements is not altered. We can conclude that the merging of the two parental genomes in this Prunus persica x Prunus dulcis hybrid does not result in a “genomic shock” with significant changes in the DNA methylation or in the transcription. The absence of major changes may facilitate using interspecific peach x almond crosses for peach improvement. Introduction copy number has been reported in interspecific hybrids Interspecific hybridization is a highly relevant process in of Helianthus and Aegilops [11, 12]. The activation and plant evolution and breeding, as it can result in pheno- mobilization of TEs after hybridization can induce impor- typic changes and sexual isolation and be at the origin of tant genome changes through many mechanisms [2], in new species [1]. Hybridization results in the combination line with Barbara McClintock ideas of TEs as controller of diverged and novel genes, which can have strong con- elements helping to reorganize the genome to overcome sequences on the phenotype [2]. Hybridization can also stress situations [13]. induce epigenetic changes, including changes of DNA TE activity is tightly controlled by epigenetic mecha- methylation and in the populations of small RNAs [2]. nisms and DNA methylation is the most obvious and fre- The genomic changes frequently induced by merging two quent chromatin modification associated to TE silencing different genomes can be so wide that they have been [14]. The mutation of different enzymes responsible of frequently referred to as “genomic shock” [3]. For exam- DNA methylation and chromatin modification results ple, important changes in gene expression have been in a decrease of DNA methylation and induces the observed in interspecific crosses of species of Senecio, activation of plant TEs [15]. Similarly, some biotic and Tragopogon or Gossypium [4, 5, 6]. On the other hand, abiotic stresses can result in a decrease of the DNA structural genome changes have also been reported, in methylation and can activate the TE transcription and particular the activation of transposable elements (TEs). mobilization [16]. The merging of two different genomes The transcriptional activation of TEs has been reported in in allopolyploids can also induce changes in DNA interspecific crosses of, for example, Spartina, Solanum or methylation and gene expression, with important conse- Nicotiana [7, 8, 9]. Transpositional activation of different quences on the phenotype [17]. For example, changes in TEs was also reported in rice introgression lines derived DNA methylation at the CONSTANS-LIKE2 gene have an from crosses with Zizania latifolia [10] and increases in TE impact on the flowering time in domesticated cotton Received: 22 February 2022; Accepted: 24 May 2022; Published: 26 May 2022; Corrected and Typeset: 1 August 2022 © The Author(s) 2022. Published by Oxford University Press on behalf of Nanjing Agricultural University. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 2 | Horticulture Research, 2022, 9: uhac127 allotetraploid species [18], and changes in histone Therefore, we performed whole genome bisulphite modifications in Arabidopsis hybrids and allopolyploids sequencing of DNA extracted from leaves of almond, results in an increased biomass, vigour and in starch peach and an almond x peach F1 hybrid. In both peach accumulation [19]. Massive changes in DNA methylation and almond, more than half of the annotated TEs were in TEs have been observed in newly formed hybrids, as, covered by mapped reads in, at least, 25% of their for example, in wheat allohexaploid [20]. A decrease length (59% in almond and 65% in peach) and were in 24-nt small RNAs, which are responsible for DNA included in downstream analyses (Table 1). In order to methylation, has been shown in F1 allopolyploids analyze the methylation of TEs in a peach x almond between Arabidopsis thaliana and Arabidopsis arenosa and hybrid we mapped the reads from the hybrid to both in intraspecific hybrids of Arabidopsis thaliana [21, 22]. parental genomes independently. The sequence of the Therefore, the merging of two genomes can modify genomes of peach and almond is highly similar, with the epigenetic silencing of TEs and frequently results an average identity of 97.99% (20 SNPs/Kb) in regions in TE activation that can induce further changes in aligning 1:1 [29]. We used mapping parameters allowing the genome. However, examples where the merging of the cross-hybridization of most reads from each one two different genomes does not result in changes in TE of the two subgenomes of the hybrid to each of the activity and genome structure have also been reported, parental genomes (see material and methods section). for example, in crosses between Arabidopsis thaliana When the reads of the hybrid were mapped onto the and Arabidopsis lyrata [23]. TE proliferation has also almond reference genome, 39% of the TEs were covered been reported to be rare in natural Helianthus hybrids, by uniquely mapped reads on at least the 25% of despite their widespread transcriptional activity [24]. The their length, whereas this percentage was higher (50%) reasons for this unpredictable outcome of the merging when they were mapped onto the peach genome. A of two different genomes, and in particular, on TE activa- global analysis of TE methylation in both peach and tion, are not known but it could be related to the level of almond and in their hybrid shows that the patterns genome divergence between the two progenitors [2, 7]. of DNA methylation are in general similar (Figure 1). Peach (Prunus persica) is one of the best-characterized As commonly found in plants [32] the highest mean species among the family Rosaceae and an important methylation was observed in CG context in both the stone fruit crop [25]. Peach does not have a functional parents and in the hybrid (about 90%, in average), gametophytic self-incompatibility system and mainly followed by the CHG context (about 60%, in average) and behaves as self-pollinating, and consequently, it shows the CHH context (about 15%, in average). A comparison of low levels of genetic diversity [26]. For this reason, the levels of methylation in peach and almond genomes breeders have started to explore the possibility to use shows a small although significant demethylation in other Prunus species as an additional source of variability the hybrid in all contexts in comparison to the two [27]. Almond (Prunus dulcis) is one of the closest species parental except in the comparison to almond in CHH to peach, both belonging to the subgenus Amygdalus [28]. context (Figure 1). The levels of DNA methylation in Peach and almond are diploid (2n = 2x = 16) and have the hybrid were compared to both peach and almond relatively small genomes (about 300 Mbp) which has independently. In all the cases in the parental genomes been sequenced [25, 29]. The two genomes show a high the TEs showed a higher level of methylation than in level of similarity and are mainly syntenic [29]. Most the hybrid, but the differences were always small, with a almond cultivars are self-incompatible and the almond maximum of 4.6% when comparing to peach genome genome is seven times more variable than peach [30]. (CHG context) and 0.7% when comparing to almond Peach and almond can be crossed to produce hybrids genome (CHG context) (Figure 1). that are frequently fertile [31]. In consequence, almond We then analysed separately the four most abundant has been considered as an interesting source of novel classes or TEs present in peach and almond [29] alleles for peach breeding [27]. Here we investigated to (Figure 2): LTR-retrotransposons, LINEs, MITEs and TIRs. what extend the crosses of peach and almond result in In general, the levels of methylation comparing the the activation of TEs that could lead to a “genomic shock”. hybrid with the parental genomes are very similar with a maximum difference of 5,1% (TIR, Peach vs Hybrid, CHG). The levels of methylation for LTR retrotransposons Results in the three genomes are similar to the ones obtained Analysis of the potential changes in the DNA for the total TEs (Figure 1). This result was expected methylation of the transposable elements in the due that LTR retrotransposons represent most of the peach x almond hybrid genome space occupied by TEs. The methylation levels TE activity is tightly controlled by epigenetic mecha- of the hybrid are similar to the two parentals except nisms [14] and DNA methylation is the most obvious in the case of CHG which is 4,8% higher in peach and frequent chromatin modification associated to TE than in the hybrid. LINE methylation shows a bimodal silencing [15]. In consequence, we decided to analyse the distribution, with average levels smaller than in the TE methylation levels in a peach x almond F1 hybrid and LTR retrotransposons, especially in the CG and CHG compare them with that of the two parental genomes. contexts. In the case of the MITEs, the average levels de Tomás et al. | 3 Table 1. Percentage of TE elements covered in at least 25% of their annotation in the DNA methylation analyses. In the hybrid, the TE elements were compared to the almond and peach genomes separately Peach Almond Hybrid (Peach ref.) Hybrid (Almond ref.) Total 65 59 50 39 LTR retro. 68 61 55 42 LINE 92 82 83 69 MITE 52 53 35 31 TIR 66 55 52 37 Figure 1. Distribution of the DNA methylation levels in the transposable elements. Violin plots for the distribution of mean methylation levels (expressed in percentage) for different methylation sites (CG, CHG and CHH) in transposable elements of the genomes of almond (A, in green), peach (P, in orange) and the interspecific hybrid (H, in blue). H (A) and H (P) represent the methylation levels of the hybrid computed using Almond or Peach genome, respectively. Statistical significance of the differences between parentals and the hybrid is presented (Wilcoxon test p-value), along with the mean methylation differences (average % of methylation in the hybrid minus average % of methylation in the parental). The numbers in the bottom represent the total number of elements analysed in each case. of methylation were similar to the LTR retrotransposons in the hybrid (5,1% in the CHG context). In conclusion, in the CG and CHG contexts, but significantly higher we observed differences in the methylation levels of the in CHH. This is similar to what has been previously different classes of TEs but only minor differences when found in different rice tissues [33]. When comparing comparing the genomes of the hybrid with the genomes the hybrid with the parental genomes, we observed that of the two parentals. In spite of these minor differences, MITEs in peach are more methylated than in the hybrid the methylation profile of the hybrid was more similar in the CG and CHG contexts, but equally methylated to the almond parental, with only a few exceptions such in CHH. No significant differences were found when as LINEs in CHG or MITEs in CHH context. comparing MITE methylation in almond and the hybrid in any of the three contexts. Finally, the average levels of Analysis of the differentially methylated regions methylation in TIRs are similar to LTR retrotransposons in the peach x almond interspecific hybrid with only a lower overall methylation in the CHG context The absence of major differences in DNA methylation of about 10–15%. Comparing the levels in the hybrid between parental genomes and the hybrid does not dis- with the level in almond, they are very similar. The TIR card the existence of local specific differences that could methylation levels in peach are a slightly higher than affect the transcriptional activities of some specific TEs. 4 | Horticulture Research, 2022, 9: uhac127 Figure 2. Distribution of the DNA methylation levels in the transposable elements separated by categories. Violin plots for the distribution of mean methylation levels (expressed in percentage) for different methylation sites (CG, CHG and CHH) in different classes of transposable elements separately (LTR retrotransposons, LINEs, MITEs and TIRs) in the genomes of almond (A, in green), peach (P, in orange) and the hybrid (H, in blue). H (A) and H (P) represent the methylation levels of the hybrid computed using Almond or Peach genome, respectively. Statistical significance of the differences between parentals and the hybrid is presented (Wilcoxon test p-value). The numbers in the bottom represent the total number of elements analysed in each case. We therefore decided to analyse the possible presence Analysis of the potential changes in the of differentially methylated regions (DMRs) between the transcription of the transposable elements in the parental lines and the hybrid. We quantified the presence peach x almond hybrid of DMRs in genes, the upstream regions of genes and in To study the potential activation of TEs by the inter- LTR retrotransposons comparing the hybrid with the two specific cross, we performed an RNA-seq analysis of parental genomes (Table 2). We concentrated on LTR- the expression in mature leaves of almond, peach and retrotransposons, as these are the most prevalent TEs in the hybrid. We analysed the possible expression of LTR peach and almond genomes [29]. In all the cases, the per- retrotransposons, LINEs and TIRs (Figure 4A). We found centages of DMRs with higher methylation in the hybrid significant transcription levels for 47 TE families: 13 respect to the parents are always around 50%, with a LTR retrotransposon, 17 LINE and 17 TIR. Among them, maximum difference of 59% in LTR retrotransposons. we found significant differential transcription between When we concentrate in the LTR retrotransposon the almond, peach and/or the hybrid in 32 families families with a higher number of DMRs (Figure 3) we (Figure 4A and B): 11 LTR retrotransposon, 12 LINE and found that in some cases most of the DMRs of the LTR- 9 TIR. In most of the cases, the differential expression is retrotransposons family correspond to demethylation in due to differences between peach and almond and the the hybrid. This is particularly true in the case of the expression in the hybrid is intermediate. In 18 families CHG context when comparing peach with the hybrid the expression was significantly higher in almond than (Figure 3A), but also, in minor intensity, in the CHH in peach, and in 13 was the opposite. Only one of the context (Figure 3B) or when comparing almond with the TE families showed lower significant expression in the hybrid (Supplementary Data 1). hybrid than in the two parental (TIR_3706) and none de Tomás et al. | 5 Table 2. DMRs found when comparing the hybrid genome with the parental almond and genomes. We distinguish between the three different targets of methylation (CG, CHG and CHH), three different regions of the genome (LTR retrotransposons, genes and upstream genes) and those DMRs showing higher methylation in the hybrid (Hybrid > Almond; Hybrid > peach) and showing higher methylation in the parental genome (Almond > Hybrid; Peach > Hybrid). Genes means the region between the ATG and the stop codon (including introns). Upstream genes means the 1000 bp upstream the ATG Hybrid vs Almond CG CHG CHH Total Hybrid Almond Total Hybrid Almond Total Hybrid Almond > > > > > > Almond Hybrid Almond Hybrid Almond Hybrid LTR retro. 37 41% 59% 209 49% 51% 596 59% 41% Genes 837 54% 46% 527 52% 48% 978 53% 47% Upstream genes 131 46% 54% 154 50% 50% 434 52% 48% Hybrid vs Peach CG CHG CHH Total Hybrid Peach Total Hybrid Peach Total Hybrid Peach > > > > > > Peach Hybrid Peach Hybrid Peach Hybrid LTR retro. 41 41% 59% 765 45% 55% 722 52% 48% Genes 1319 52% 48% 697 56% 44% 839 53% 47% Upstream genes 254 55% 45% 235 54% 46% 672 54% 46% Figure 3. Differentially methylated regions in LTR retrotransposons when comparing peach and hybrid genomes. Methylation levels of DMRs in the LTR retrotransposon families comparing peach with the hybrid in CHG (A) and CHH (B) contexts. Blue dots indicate that the DMR is more methylated in the hybrid. Orange dots indicate that the DMR is more methylated in peach. Only families with at least 5 DMRs are shown. was expressed at a higher level than in the two parental species with higher levels of transcription is the one with species. higher copy number and, in the cases in which that does A more detailed analysis of the transcribed LTR retro- not happen (LTR_48 and LTR_124), the copy numbers are transposon families showed that in one case (LTR_99) very similar. On the other hand, all the transcribed LTR only the almond genome contains full-length copies and retrotransposon families except one (LTR_138) contain this is correlated with a higher expression in almond relatively young copies with an estimated age of 1,4 Mya (Table 3). In 8 of the other 10 transcribed families the or less. In all the families, the parental species with the 6 | Horticulture Research, 2022, 9: uhac127 Table 3. Transcribed LTR retrotransposon families. A, almond; P, peach; H, hybrid Family Type Pattern of transcription Number of full copies Minimal estimated time of insertion (Mya) Almond Peach Almond Peach 3 Gypsy A < (P = H) 11 27 1.7 0.0 28 Gypsy (A = H) > P 16 10 1.7 0.6 48 Copia A < (P = H) 26 23 0.0 0.0 53 Copia A>H>P 26 12 0.0 0.0 81 Copia A < (P=H) 7 8 1.4 1.4 88 Copia A > (H = P) 10 5 0.8 0.8 99 Copia (A = H) > P 3 0 1.4 - 108 Gypsy A>H>P 14 6 0.0 0.5 124 Unclassified A>H>P 3 4 10.4 4.3 138 Unclassified A (H = P) 5 2 1.3 4.6 most of the LTR retrotransposon families with differen- tial expression several copies are expressed but from four of them we were able to determine a single copy responsible of all, or most, of the transcription, being in all cases recent insertions present close to genes or inside a gene. In the family LTR_3 the expressed copy is in chromosome 6 of peach (27139890–27 148 350), is estimated to be 1,8 Mya old and is located inside a gene (in an intron) in the same orientation. In the family LTR_81 the expressed copy is in the chromosome 3 of peach (19702391–19 706 564), is estimated to be 1,9 Mya old and is located 3,2 kB from the closest gene. In the family LTR_124 the expressed copy is in the chromosome 7 of almond (21257163–21 260 133), is estimated to be 10,4 Mya old and is located inside a gene (in an intron) in the same orientation. Finally, in the family LTR_141 the expressed copy is in the chromosome 8 of almond (10143038–10 147 556), is estimated to be 1,3 Mya old and is located 35 bases downstream a gene but in opposite orientation. All these data suggest that the merging of the genomes of peach and almond in a hybrid does not greatly dereg- ulate the expression of TEs and that the differential Figure 4. Transcriptomic analysis of transposable elements in leaves expression of TEs between these two genomes is mainly of almond, peach and the hybrid. A) Patterns of transcription of the TE due to the presence or absence of transcriptional active families. Vertical higher position indicates more transcription (P, peach; H, hybrid; A, almond). Total means the total number of families in peach copies in each of them. and almond genomes containing at least one full-length element. B) Heat-map of the transcription levels of the TE families showing Analysis of the potential changes in gene expression in at least one of the genotypes. Higher expression is transcription in the peach x almond hybrid indicated in red and lower expression in dark blue. We show the results of three replicates per sample. We analysed the possible changes in gene expression in the hybrid (Figure 6). As already mentioned, the peach and almond genome show a high degree of sequence younger element is the one showing higher transcription identity (mean of 97.99% in regions aligning 1:1, [27]. (Table 3). Therefore, to facilitate the comparison of the level To validate the RNA-seq data we performed qRT-PCR of expression, we decided to map the RNA-seq reads analysis for nine of the LTR retrotransposon families from peach, almond and the hybrid to a single gene differentially expressed and the results confirmed their model dataset, that of peach or, alternatively, that of expression profile showing a general agreement with the almond. A global comparison of the expression levels transcript abundance estimated by RNA-seq (Figure 5). of 13 620 genes orthologous between peach and almond Next, we tried to identify which of the copies of the (see methods for parameters used to define orthology), differentially transcribed LTR retrotransposon families showed an almost perfect correlation between the two were the ones that produce transcripts in leaves. For deduced profiles (Pearson correlation coefficient = 0.99) de Tomás et al. | 7 Figure 5. Transcription of LTR retrotransposon families in leaves of almond, peach and the hybrid. Transcription levels of LTR retrotransposons in leaves of almond (A, green), peach (P, orange) and the F1 hybrid (H, blue) analysed by qRT-PCR and RNA-seq. The LTR retrotransposon family is indicated in the top. (Supplementary Data S2). We present here the results in the production of sRNAs, in the defence against of the analysis performed using the peach gene models viruses [35] and in gene regulation and DNA methylation (Figure 6) and the almond gene models (Supplementary [36]. We next analysed other genes possibly involved in Data 3). From the 26 873 genes present in the annotated DNA methylation encoding DNA methyltransferases or peach genome [34] we found that 22 274 of them are DNA demethylase (Table 4) but none of them showed significantly expressed in at least one of the three differential expression among peach, almond and the genotypes (peach, almond and hybrid). For 17 439 genes hybrid. (78,3%) the levels of transcription were similar in the three genotypes. For 2389 genes (10,7%) the expression was higher in peach respect to almond and in the F1 Discussion hybrid the expression was intermediate or similar to one Hybridization is a very relevant and relatively frequent of the two parents. For 2234 genes (10,0%) the expression process in plant evolution [2], that has also been used in was higher in almond respect to peach and in the F1 plant breeding. For example, crosses of different varieties hybrid was intermediate or similar to one of the two to produce hybrids presenting superior phenotypes as parents. In 152 genes (0,7%) the expression was higher compared with their parents (hybrid vigour) is widely in the hybrid than in the two parental and in 60 genes used in crops [37] and interspecific hybridization with (0,3%) the expression was lower in the hybrid. crop wild relatives, or highly variable related species, is Among the genes whose expression is significantly frequently used to expand the species variability used lower in the hybrid there are two genes, Prupe.1G332600.1 for breeding [38]. Peach is a self-fertile and naturally self- and Prupe.1G334500.1 (Table 4), annotated as poten- pollinating species with very low genetic variability [39]. tially encoding an “RNA-dependent RNA polymerase, The use in breeding programs of interspecific crosses eukaryotic-type” showing sequence similarity to the Ara- with other species of the genus Prunus has been histor- bidopsis RDR1 (AT1G14790), one of the enzymes involved ically used to increase peach genetic variability [27]. In 8 | Horticulture Research, 2022, 9: uhac127 Table 4. Transcription of genes encoding proteins possibly involved in DNA methylation Almond Hybrid Peach rlog value SD rlog value SD rlog value SD RNA-dependent RNA polymerase, eukaryotic-type Prupe.1G332600.1 5,89 0,34 5,09 0,16 5,57 0,42 Prupe.1G334500.1 6,11 0,43 5,45 0,19 6,04 0,28 DNA methyltransferase Prupe.7G183100.1 7,08 0,30 6,64 0,19 6,67 0,30 Prupe.6G011600.1 2,70 0,03 2,71 0,06 2,96 0,15 Prupe.8G038800.1 9,62 0,33 9,21 0,15 8,97 0,15 Prupe.6G322700.1 10,89 0,10 10,90 0,14 10,91 0,26 DNA demethylase Prupe.7G118000.1 11,37 0,30 11,36 0,13 11,26 0,09 Prupe.7G005000.1 12,34 0,13 12,41 0,15 12,11 0,27 Prupe.6G119100.1 6,48 0,37 6,63 0,10 7,02 0,09 changes that can be so widespread that have been defined as a “genomic shock” [3]. Here we analysed the possible genetic and epigenetic changes associated with the crossing of peach and almond to produce an interspecific hybrid. Transposons are the primary target of epigenetic mechanisms, and DNA methylation is the main epigenetic modification associated with TE silencing [15]. Therefore, we have compared the DNA methylation of TEs in peach and almond with that of a F1 interspecific hybrid. Our results show that there are no major differences in the methylation of TEs between the two parental species or between both parents and the hybrid. We found some differentially methylated regions that overlap with LTR retrotransposons, the main order of TEs in peach and almond [29]. In some cases, as for the CHG context, and in particularly when comparing the hybrid with peach, most of the DMRs of the same TE family are demethylated in the hybrid, suggesting a possible weakening of the epigenetic silencing and an increased potential for activation associated with the interspecific cross. However, this demethylated trend has no parallel for the CHH context, where different copies of the same LTR retrotransposon family can show hypermethylated and hypomethylated DMRs in the hybrid. Moreover, the analysis of the transcription of the Figure 6. Transcription of genes in leaves of almond, peach and the LTR retrotransposons in leaves in the two parents and in hybrid. Transcriptomic analysis of the genes in leaves of almond, peach and the hybrid. A) Patterns of transcription of the genes. Vertical higher the hybrid did not show the reactivation of any of the LTR position indicates more transcription (P, peach; H, hybrid; A, almond). retrotransposon families after hybridization. In addition, B) Average transcription levels of the genes showing significant we neither found transcriptional activation of other types differential expression in peach, the F1 hybrid and almond. Higher expression is indicated in red and lower expression in dark blue. of TEs in leaves of the hybrid. These results suggest that the cross of peach and almond did not result in important changes in the regulation of TEs in general and in the LTR the last years a growing interest has emerged in the use retrotransposons in particular. Among the genes that of these related species for peach breeding, mainly as a show a reduced expression in the hybrid there are two source of pathogen resistances [40]. Almond has become genes potentially encoding for an RNA-dependent RNA an interesting choice for introgressing new genes into polymerase showing similarity with the RDR1 protein peach, mainly due to the high genetic variability present from Arabidopsis, which is involved in the production in almond and also for fruit quality traits [27]. of sRNAs, viral defence and DNA methylation [35, 36]. In addition to the combination of diverged alleles and However, a closer look at more genes involved in DNA different genes, the merging of two different genomes methylation dynamics [41] did not reveal any difference can also be accompanied by epigenetic and structural of expression. Only 1% of the genes showed transgressive de Tomás et al. | 9 expression in the hybrid which reinforces the idea that Gene and TE reference datasets no major genomic changes are induced by the merging Gene annotations were obtained from Genome Database of the peach and almond genomes in the hybrid. for Rosaceae (GDR; https://www.rosaceae.org/). TE library There are many examples were interspecific crosses described in [29] was curated to retain only high- result in genome demethylation and/or TE activation confidence TE annotations (based on the presence of [11, 12]. In consequence, the lack of signs of a “genomic structural features, coding domains or homology to shock” in the peach x almond hybrid may seem surpris- known TEs), resulting in a more stringent annotation ing. However, not always interspecific crosses result in (Supplementary Data S4, S5 and S6). genome demethylation and/or TE activation as it has been shown, for example, in crosses between Arabidopsis Analysis of DNA methylation thaliana and Arabidopsis lyrata [23]. It has been proposed Library preparation was carried out [29], including two that when merging two different genomes in a hybrid bisulfite conversion rounds to ensure high conversion the intensity of the genome rearrangement and TE mobi- rate. We calculated bisulfite conversion rates by mapping lization could depend on the TE load imbalance and the reads to the non-methylated chloroplast genome. The phylogenetic distance between the parents [7]. Almond conversion rates were 97% (almond), 98% (hybrid) and and peach are Prunus species of the same subgenera, 98% (peach). After processing for removal of adapter Amygdalus, and have diverged only six million years ago and low-quality sequences, we obtained between 41.0 [29]. Considering that the mean generation time for these and 44.5 million reads for each of the two independent species is 10 years, this explains the high conservation of replicates of the three genotypes. Almond and peach their genome sequence which is as low as 20 nucleotide trimmed reads were aligned to the corresponding substitutions per Kbp [29]. In addition, the two genomes parental genomes. Reads from the hybrid were aligned are also very similar in the proportion and types of TEs independently to peach and to almond genomes. they content, sharing the majority of TE families and Cytosine methylation was analysed using Bismark many individual TE insertions [29]. Therefore, the small v.0.19.1 [44] and SeqMonk software package (v.1.41; Bab- phylogenetic distance between peach and almond and raham Institute; http://www.bioinformatics.babraham. their shared TE load could be the reason for the absence ac.uk/projects/seqmonk/). We only included cytosine of a detectable “genomic shock” associated with their positions that had been sequenced at least three interspecific cross. times. Reads were mapped with Bismark keeping only In conclusion, our work shows that the merging of unique best alignments. A read was considered to peach and almond genomes in an interspecific hybrid is align uniquely if one alignment exists that had fewer not associated with TE reactivation or general alterations mismatches to the genome than any other alignment in DNA methylation levels and has not a major impact on (or if there was no other alignment). We used a slightly gene expression. The absence of alterations in the hybrid relaxed mapping threshold (Bismark parameters: -N may facilitate the use of almond as a source of new 1, —score_min L,0,-0.6) to allow the reads from both genetic variability for breeding the low variable peach hybrid subgenomes to map to either peach or almond species. reference genomes. We used SeqMonk to quantify DNA methylation levels. SeqMonk filters each call position by the degree of coverage to produce weighted methy- Materials and methods lation values. We focused only on TE copies covered Plant material and growth conditions by sequencing reads in at least 25% of their length Leaves of Prunus dulcis cv Texas, Prunus persica cv Early (Supplementary Data S7). Gold and one interspecific F1 hybrid MB 1.37 were To identify differentially methylated regions (DMRs) collected from the Experimental Station of Lleida in we used the R package DMRCALLER [45] using a sliding Gimenells (Catalonia, Spain) kindly provided by IRTA [27]. window of 50 bp and a 50 bp sliding interval. P-value for They were cultivated in the field. Fully expanded leaves each window was calculated using Fisher’s exact test. were collected at the end of September from normally The p-values were adjusted for multiple testing using watered trees at the same time of the day. Seven leaves Benjamini and Hochberg’s method to control the false per genotype were harvested from three replicates per discovery rate [46]. Bins with fewer than three cytosines genotype. in the specified context or < 0.25 difference in methyla- tion proportion between the two conditions or an average DNA and RNA isolation number of reads lower than 8 were discarded. Finally, High molecular weight genomic DNA was isolated using bins that were at less 300 bp were joined. a sorbitol pre-wash [42] followed by an adapted CTAB method [43]. Total RNA was extracted using the Maxwell RNA sequencing and analysis RSC Plant RNA Kit and the Maxwell RSC instrument The RNA-seq libraries were obtained from 2–4 μg of (Promega Corporation, Madison, WI, USA). Complete total RNA. For each parental and hybrid genotype, DNA removal was obtained using the DNA-free DNA three biological replicates were collected. The RNA-seq Removal Kit (Invitrogen™, Carlsbad, CA, USA). libraries were produced using the Truseq stranded mRNA 10 | Horticulture Research, 2022, 9: uhac127 protocol and were sequenced on Illumina platform Acknownledgements NextSeq 500 (2x150 bp, Paired-end). To analyse TE We thank Pere Arús, Maria José Aranzana and Iban transcription, we used the methodology described in [47] Eduardo for their help in providing us with the plant with minor modifications. The most similar genomic material and useful discussions. Raúl Castanera holds a copy to each assembled contig was identified as a Juan de la Cierva Incorporación Postdoctoral fellowship family representative (Supplementary Data S8), using IJC2020–045949-I funded by MCIN/AEI/10.13039/501100 a length coverage cut-off of 80% for retrotransposons 011033 and by the “European Union NextGenera- and 40% for DNA transposons. RNA-seq reads were tionEU/PRTR”. This work was funded by grants from the aligned to peach transcript models concatenated with Spanish Ministerio de Economía, Industria y Competi- peach and almond representative TEs using Bowtie2 tividad (AGL2016–78992-R/FEDER), Ministerio de Ciencia [48] through the RSEM package in default parame- e Innovación (PID2019-106374RB-I00/AEI/10.13039/5011 ters, keeping only reads aligned in the sense strand. 00011033), by “ERDF A way of making Europe” and by the Differential expression analysis was performed using CERCA Programme of the Generalitat de Catalunya. We DESeq2 with a Log fold-change cut-off of one. False also acknowledge financial support from the Spanish Discovery Rate (FDR) was used for multiple-testing Ministerio de Economía y Competitividad through the correction. DESeq2 regularized log (rlog) values were “Severo Ochoa Programme for Centres of Excellence in used as normalized expression data for heatmap R&D” 2016–2019 (SEV-2015-0533) and CEX2019–000902-S construction. funded by MCIN/AEI/10.13039/501100011033. Validation of retrotransposon differential transcription Author contributions Quantitative RT-PCR analyses were performed using CdT performed the transcriptomic analyses. AB per- three independent RNA extractions per genotype. The formed the DNA methylation analyses. RC contributed cDNAs were synthesized using SuperScript® III Reverse to the bioinformatic analyses. JMC and CMV conceived Transcriptase (Invitrogen™, Carlsbad, CA, USA). The the research and designed the experiments. All authors primers for RT-PCR are listed in Supplementary Data S7. contributed to the final version of the manuscript. The qRT-PCR were performed in a Roche LightCycler II using Roche’s SYBR green Master Mix (Roche Applied Science) with the initial denaturation step of 5 min at Data availability 95◦ C, followed by 40 RT-PCR cycles (10 s at 95◦ C, 10 s All data used in this study are publicly available. Raw and at 56◦ C, and 10 s at 72◦ C). The Translation Elongation processed RNA-seq and bisulfite sequencing data has Factor (TEF2) and the RNA Polymerase II (RPII) were used been deposited to NCBI GEO repository under the acces- as internal controls to normalize the expression of the sion GSE198152. Genome sequences and annotations are tested LTR retrotransposons. The relative levels of gene available in the Genome Database for Rosaceae (https:// expression were calculated using the 2 − Ct method. www.rosaceae.org/). The specificity of the primers and their product length were verified by agarose gel electrophoresis. The primers for qRT-PCR are listed in Supplementary Data S9. Conflict of interests The identification of the expressed TE copies was performed using RT-PCR using different sets of primers All authors declare no conflict of interests. (Supplementary Data S7). 1 μg of total RNA was reverse transcribed using SuperScript® III Reverse Transcriptase Supplementary data (Invitrogen, Carlsbad, CA, USA). PCR cycling conditions were 2 min at 95◦ C, followed by 35 cycles of 95◦ C for Supplementary data is available at Horticulture Research 30 s, 30 s at the annealing Tm and 1 min/kb at 72◦ C, online. and a final step of 10 min at 72◦ C. Negative reverse transcriptase and non-template controls were used. References PCR results were observed using a 1% (w/v) agarose gel. PCR fragments were extracted using Macherey-Nagel™ 1. Mason AS, Batley J. Creating new interspecific hybrid and poly- NucleoSpin™ Gel and PCR Clean-up Kit (Fisher Scientific, ploid crops. Trends Biotechnol. 2015;33:436–41. 2. Nieto Feliner G, Casacuberta J, Wendel JF. Genomics of UK), cloned into the pGEM-Teasy plasmid using pGEM® -T evolutionary novelty in hybrids and Polyploids. Front Genet. Easy Vector Systems kit (Promega Corporation, Madison, 2020;11:792. WI, USA), introduced into E.coli and amplified. 8 colonies 3. Comai L, Madlung A, Josefsson C et al. Do the different parental were selected per PCR fragment and their inserts were ‘heteromes’ cause genomic shock in newly formed allopoly- sequenced. Sequences were compared with the parental ploids? Philos Trans R Soc Lond Ser B Biol Sci. 2003;358:1149–55. genomes and the expressed copies were identified only if 4. Buggs RJ, Doust AN, Tate JA et al. Gene loss and silencing the amplified sequences were more than 99,5% identical in Tragopogon miscellus (Asteraceae): comparison of natural and to the genomic sequence. synthetic allotetraploids. Heredity. 2009;103:73–81. de Tomás et al. | 11 5. Hegarty MJ, Barker GL, Wilson ID et al. Transcriptome shock after 26. Donoso JM, Eduardo I, Picañol R et al. High-density mapping interspecific hybridization in senecio is ameliorated by genome suggests cytoplasmic male sterility with two restorer genes in duplication. Curr Biol. 2006;16:1652–9. almond × peach progenies. Hortic Res. 2015;2:15016. 6. Yoo MJ, Szadkowski E, Wendel JF. Homoeolog expression bias 27. Donoso JM, Picañol R, Serra O et al. Exploring almond genetic and expression level dominance in allopolyploid cotton. Heredity. variability useful for peach improvement: mapping major genes 2013;110:171–80. and QTLs in two inter-specific almond • peach populations. Mol 7. Mhiri C, Parisod C, Daniel J et al. Parental transposable element Breed. 2016;36:16. loads influence their dynamics in young Nicotiana hybrids and 28. Hodel RGJ, Zimmer E, Wen J. A phylogenomic approach allotetraploids. New Phytol. 2019;221:1619–33. resolves the backbone of Prunus (Rosaceae) and identifies sig- 8. Parisod C, Salmon A, Zerjal T et al. Rapid structural and epige- nals of hybridization and allopolyploidy. Mol Phylogenet Evol. netic reorganization near transposable elements in hybrid and 2021;160:107118. allopolyploid genomes in Spartina. New Phytol. 2009;184:1003–15. 29. Alioto T, Alexiou KG, Bardil A et al. Transposons played a major 9. Paz RC, Rendina González AP, Ferrer MS et al. Short-term role in the diversification between the closely related almond hybridisation activates Tnt1 and Tto1 Copia retrotransposons and peach genomes: results from the almond genome sequence. in wild tuber-bearing Solanum species. Plant Biol. 2015;17:860–9. Plant J. 2020;101:455–72. 10. Wang HY, Tian Q, Ma YQ et al. Transpositional reactivation of 30. Velasco D, Hough J, Aradhya M et al. Evolutionary genomics of two LTR retrotransposons in rice-Zizania recombinant inbred peach and almond domestication. G3 (Bethesda). 2016;6:3985–93. lines (RILs). Hereditas. 2010;147:264–77. 31. Marimon N, Luque J, Arús P et al. Fine mapping and iden- 11. Senerchia N, Felber F, Parisod C. Genome reorganization in F1 tification of candidate genes for the peach powdery mildew hybrids uncovers the role of retrotransposons in reproductive resistance gene Vr3. Hortic Res. 2020;7:175. isolation. Proc R Soc B Biol Sci. 2015;282:20142874. 32. Zakrzewski F, Schmidt M, Van Lijsebettens M et al. DNA methy- 12. Ungerer MC, Strakosh SC, Zhen Y. Genome expansion in three lation of retrotransposons, DNA transposons and genes in sugar hybrid sunflower species is associated with retrotransposon beet (Beta vulgaris L.). Plant J. 2017;90:1156–75. proliferation. Curr Biol. 2006;16:R872. 33. Zemach A, Kim MY, Silva P et al. Local DNA hypomethylation 13. McClintock B. The significance of responses of the genome to activates genes in rice endosperm. Proc Natl Acad Sci U S A. challenge. Science. 1984;226:792–801. 2010;107:18729–34. 14. Fultz D, Slotkin RK. Exogenous transposable elements cir- 34. Verde I, Jenkins J, Dondini L et al. The peach v2.0 release: high- cumvent identity-based silencing, permitting the dissection of resolution linkage mapping and deep resequencing improve expression-dependent silencing. Plant Cell. 2017;29:360–76. chromosome-scale assembly and contiguity. BMC Genomics. 15. Deniz Ö, Frost JM, Branco MR. Regulation of transposable ele- 2017;18:225. ments by DNA modifications. Nat Rev Genet. 2019;20:417–31. 35. Leibman D, Kravchik M, Wolf D et al. Differential expression 16. Gutzat R, Mittelsten Scheid O. Epigenetic responses to stress: of cucumber RNA-dependent RNA polymerase 1 genes during triple defense? Curr Opin Plant Biol. 2012;15:568–73. antiviral defence and resistance. Mol Plant Pathol. 2018;19:300–12. 17. Ding M, Chen ZJ. Epigenetic perspectives on the evolution and 36. Wang N, Zhang D, Wang Z et al. Mutation of the RDR1 gene domestication of polyploid plant and crops. Curr Opin Plant Biol. caused genome-wide changes in gene expression, regional vari- 2018;42:37–48. ation in small RNA clusters and localized alteration in DNA 18. Song Q, Zhang T, Stelly DM et al. Epigenomic and functional methylation in rice. BMC Plant Biol. 2014;14:177. analyses reveal roles of epialleles in the loss of photoperiod sen- 37. Rajendrakumar P, Hariprasanna K, Seetharama N. Prediction of sitivity during domestication of allotetraploid cottons. Genome Heterosis in crop plants-status and prospects. Am J Exp Agric. Biol. 2017;18:99. 2015;9:1–16. 19. Ni Z, Kim ED, Ha M et al. Altered circadian rhythms regulate 38. Warschefsky E, Penmetsa RV, Cook DR et al. Back to the wilds: growth vigour in hybrids and allopolyploids. Nature. 2009;457: tapping evolutionary adaptations for resilient crops through 327–31. systematic hybridization with crop wild relatives. Am J Bot. 20. Yaakov B, Kashkush K. Massive alterations of the methylation 2014;101:1791–800. patterns around DNA transposons in the first four generations 39. Micheletti D, Dettori MT, Micali S et al. Whole-genome anal- of a newly formed wheat allohexaploid. Genome. 2011;54:42–9. ysis of diversity and SNP-major gene association in peach 21. Groszmann M, Greaves IK, Albertyn ZI et al. Changes in 24-nt Germplasm. PLoS One. 2015;10:e0136803. siRNA levels in Arabidopsis hybrids suggest an epigenetic contri- 40. Martínez-Gómez P, Rubio M, Dicenta F et al. Resistance to plum bution to hybrid vigor. Proc Natl Acad Sci U S A. 2011;108:2617–22. pox virus (RB3.30 isolate) in a group of California almonds and 22. Ha M, Lu J, Tian L et al. Small RNAs serve as a genetic buffer transfer of resistance to peach. J Am Soc Hortic Sci. 2004;129: against genomic shock in Arabidopsis interspecific hybrids and 544–8. allopolyploids. Proc Natl Acad Sci U S A. 2009;106:17835–40. 41. Rothkegel K, Espinoza A, Sanhueza D et al. Identification of 23. Göbel U, Arce A, He F et al. Robustness of transposable element DNA methylation and Transcriptomic profiles associated with regulation but no genomic shock observed in interspecific Ara- fruit Mealiness in Prunus persica (L.) Batsch. Front Plant Sci. bidopsis hybrids. Genome Biol Evol. 2018;10:1403–15. 2021;12:684130. 24. Kawakami T, Dhakal P, Katterhenry AN et al. Transposable 42. Inglis PW, Pappas MCR, Resende LV et al. Fast and inexpen- element proliferation and genome expansion are rare in con- sive protocols for consistent extraction of high quality DNA temporary sunflower hybrid populations despite widespread and RNA from challenging plant and fungal samples for high- transcriptional activity of LTR retrotransposons. Genome Biol Evol. throughput SNP genotyping and sequencing applications. PLoS 2011;3:156–67. One. 2018;13:e0206085. 25. Arús P, Verde I, Sosinski B et al. The peach genome. Tree Genet 43. Doyle JJ, Doyle JL. Isolation of plant DNA from fresh tissue. Focus. Genomes. 2012;8:531–47. 1990;12:13–5. 12 | Horticulture Research, 2022, 9: uhac127 44. Krueger F, Andrews SR. Bismark: a flexible aligner and methyla- practical and powerful approach to multiple testing. J R Stat Soc tion caller for bisulfite-Seq applications. Bioinformatics. 2011;27: Series B Methodol. 1995;57:289–300. 1571–2. 47. Vendrell-Mir P, López-Obando M, Nogué F et al. Different fam- 45. Catoni M, Tsang JM, Greco AP et al. DMRcaller: a versatile ilies of Retrotransposons and DNA transposons are actively R/Bioconductor package for detection and visualization of dif- transcribed and may have transposed recently in Physcomitrium ferentially methylated regions in CpG and non-CpG contexts. (Physcomitrella) patens. Front Plant Sci. 2020;11:1274. Nucleic Acids Res. 2018;46:e114. 48. Langmead B, Salzberg SL. Fast gapped-read alignment with 46. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a bowtie 2. Nat Methods. 2012;9:357–9.