2018 Volume 68 Issue 2 Pages 168-176
In plants, several transposable elements are conserved across species. We found a homolog of ONSEN, which is a heat-activated retrotransposon originally isolated from Arabidopsis thaliana, in Vigna. The ONSEN-like elements (VaONS) were detected in all the analyzed Japanese accessions of Vigna angularis (adzuki bean) by Southern blot analysis. However, VaONS sequences were observed to be polymorphic in the different accessions. Interestingly, extrachromosomal DNA (ecDNA) was detected in some accessions of adzuki bean, indicating the conserved heat-activation of VaONS. Furthermore, we successfully induced retrotransposition of VaONS in adzuki plant regenerated through callus. Findings of our study should provide a new tool for molecular breeding of adzuki bean.
Transposable elements (TEs) are abundant in plant and animal genomes and can influence genome evolution and expression of genes (Kazazian 2004, Kumar and Bennetzen 1999, Wessler 1996). TEs are mainly divided into two classes based on their transposition mechanism (Finnegan 1989, Flavell et al. 1994). Class-I TEs are retrotransposons and are transposed via an RNA intermediate, by a “copy and paste” mechanism. Class-II TEs involve DNA transposons and are transposed by a “cut and paste” mechanism. Retrotransposons are divided into two families based on the presence or absence of a long terminal repeat (LTR). The LTR retrotransposons are further classified into Ty1/copia-like or Ty3/gypsy-like depending on the order of their coding domains.
Although most TEs are silenced because of epigenetic modifications, such as DNA methylation or histone modification, some TEs are activated under stress conditions (Bennetzen 1987, Chandler and Walbot 1986, Grandbastien 2004, Grandbastien et al. 1997, Hashida et al. 2003, Henderson and Jacobsen 2007, Hirayama et al. 2009, Hirochika 1993, Lisch 2009, Scortecci et al. 1997, Steward et al. 2000, Zeller et al. 2009). Some LTR retrotransposons contain cis-regulatory sequences in their 5′-LTR that can be recognized by a stress-induced transcription factor, which triggers the expression of transposon in response to a particular stimulus (Casacuberta and Gonzalez 2013). The regulatory sequences are variable, suggesting that retrotransposons could have evolved through the modification of regulatory features. Previously, we found heat stress-induced activation of a Ty1/copia-like retrotransposon, ONSEN, in Arabidopsis thaliana (Ito et al. 2011). The activation of ONSEN requires a heat-induced transcription factor, HSFA2 (Cavrak et al. 2014). HSFA2 binds to a cis-regulatory sequence (heat response element, HRE) in the promoter of ONSEN LTR. The activated ONSEN is not only transcribed but is also transposed in heat-stressed plants that are deficient in the RNA-directed DNA methylation (RdDM) pathway (Ito et al. 2011, Matsunaga et al. 2012).
Activation of TEs in response to stress induces mutations that could help the organism adapt to new environmental conditions (McClintock 1984). In A. thaliana, ONSEN is preferentially inserted within or close to genes (Ito et al. 2011) and changes the expression of the flanking genes. Transposition of ONSEN was reported to generate a mutation in an abscisic acid (ABA) responsive gene, resulting in an ABA-insensitive phenotype in A. thaliana (Ito et al. 2016). As such, ONSEN family retrotransposons are potential genetic tools for generation of novel phenotypes in host plants.
Ty1/copia-like retrotransposons are ubiquitous in legume crops (Kanazawa et al. 2009, Lall et al. 2002, Patil et al. 2015, Pearce 2007, Pearce et al. 1996, Rajput and Upadhyaya 2010a, 2010b, Sant et al. 2000, Wang et al. 2004, Xiao et al. 2007); therefore, in the present study, we studied whether the ONSEN family members are present as functional elements in adzuki bean [Vigna angularis (Willd.) Ohwi & Ohashi], an important legume crop in Japan. Adzuki bean is one of the major crops in Hokkaido. Intensive efforts are invested in producing new cultivars of this bean. However, small genetic variation in the cultivated adzuki beans has often been a bottleneck in finding the desired traits for cross breeding. Mutagenesis using TEs might be helpful in expanding the genetic diversity of this bean. In addition, it is now easy to find repeat elements in adzuki bean, because a reference-grade genome sequence has become available (Sakai et al. 2015).
In this study, in addition to the cultivated adzuki bean, we also included wild adzuki bean, because many accessions are available from NARO Genebank (http://www.gene.affrc.go.jp/index_en.php). The wild adzuki bean is considered to be a progenitor of the cultivated adzuki bean and both are highly cross-compatible with each other (Wang et al. 2004). Moreover, because the wild accessions in East Japan have a different karyotype from those in West Japan (Wang et al. 2015), we also investigated the correlation between the insertion polymorphisms of ONSEN and the geographic origins of the accessions.
Here we report the characterization of a heat-activated ONSEN family in the wild and cultivated adzuki bean. We found de novo insertions in regenerated plants, indicating that at least one ONSEN element can be fully activated in the calli.
The sequences of transposable elements that are orthologous to the ONSEN family in A. thaliana (AtONS) were identified by BLAST search against available public databases of Vigna radiata (Kang et al. 2014) and Vigna angularis (Sakai et al. 2015). The whole genome sequences were downloaded and local BLAST search was performed with the sequence of reverse transcriptase region of AtONS. The obtained sequences of the RT region were aligned and a phylogenetic tree was constructed using other COPIA family sequences from A. thaliana. The sequences that were orthologous to AtONS were retained for further analyses. To obtain full-length sequences of Vigna ONSEN-like elements, 10-kb regions on each side, flanking the RT regions, were analyzed using the LTR finder (Xu and Wang 2007). The sequences between the LTRs were aligned to determine the consensus sequence for the reference full-length sequence of each species. Because the consensus sequence from V. angularis had a 1-bp insertion, which caused a frameshift in the CDS, the 1-bp insertion was deleted to obtain a full-length reference sequence. The full-length reference sequences were 4895 and 4890 bp in length for V. angularis (413 bp LTR) and V. radiata (410 bp LTR), respectively. The reference sequences (Supplemental materials) were then used for BLAST search against the whole genome sequences to estimate the copy number in each species. The ONSEN-like elements in each species were retrieved when the BLAST hits were more than 300 bp in length and were not similar to the sequences of other COPIA family members.
A set of 117 accessions of V. angularis used for Southern blot analysis was procured from the National Institute of Agrobiological Sciences (Tsukuba, Japan). The accessions were categorized by habitat in East and West Japan (Supplemental Tables 2, 3) based on a previous study by Wang et al. (2015). The habitat of cultivated accessions was referred to the habitat of wild accessions so that some cultivated accessions were not assigned (Supplemental Table 3). One commercially bred adzuki, Tanba Dainagon, was obtained from Takii & Company Limited, Kyoto, Japan. One variety of V. angularis, named Shumari, was obtained from the Hokkaido Research Organization (Obihiro, Japan). The plants were grown on Murashige and Skoog (MS) plates at 21°C under continuous light conditions.
Genomic DNA was isolated using the Nucleon PhytoPure DNA extraction kit (GE Healthcare Life Science, Chicago, IL, USA). Southern blotting was performed as described previously (Miura et al. 2004). The DNA was digested with EcoRV and the hybridization signals were detected using a radiolabeled VaONS-specific probe (Supplemental Table 4) generated using the Megaprime DNA Labeling System (GE Healthcare Life Science), in high sodium dodecyl sulfate hybridization buffer (Church and Gilbert 1984). The independence of the detected bands was statistically tested by Fisher’s exact test.
Seven-day-old seedlings grown on MS plates at 21°C under continuous light conditions were subjected to a temperature shift from 4°C for 24 h to 40°C for 24 h. DNA and RNA were immediately extracted from the stressed seedlings after the heat treatment. As a control, seedlings were subjected to a temperature shift from 4°C for 24 h to 21°C for 24 h.
Total RNA was extracted from the whole seedlings using TRI Reagent (Sigma Aldrich, St. Louis, MO, USA), according to the manufacturer’s instructions. Approximately 3–5 μg of the total RNA was treated with RQ1 RNase-free DNase (Promega, Madison, WI, USA). For VaONS transcription, RNA was reverse-transcribed using the ReverTraAce qPCR RT Kit (Toyobo, Osaka, Japan) with a random primer. Polymerase chain reaction was performed using TaKaRA Ex Taq (TaKaRA, Shiga, Japan). The reaction conditions were as follows: 5 min at 94°C; 25 cycles of 94°C (30 sec), 55°C (30 sec), and 72°C (1 min); 7 min at 72°C. As a control, ACT2 transcript was analyzed by OneStep RT-PCR Kit (QIAGEN, Venlo, Netherlands) using the following reaction conditions: 30 min at 50°C; 15 min at 95°C; 35 cycles of 94°C (30 sec), 55°C (30 sec), and 72°C (1 min); 7 min at 72°C. The sequences of the primers used are summarized in Supplemental Table 4.
The phylogenetic relationships were analyzed using the neighbor-joining method. The synonymous distances were calculated from the reverse transcriptase region (2994 to 3761 and 2991 to 3758 nucleotide positions for reference sequences of V. angularis and V. radiata, respectively) of the ONSEN sequences by the method of Nei and Gojobori. Indel sites were analyzed with pair-wise deletion option. All the analyses were performed with MEGA 7.0 (Kumar et al. 2016). The bootstrap probabilities were estimated by 500 replications.
The seedlings of V. angularis were grown on MS plate with continuous light at 27°C for 12 days and subsequently subjected to a temperature shift from 4°C for 24 h to 40°C for 24 h. The hypocotyl of a heat-stressed plant was transferred to callus induction medium, consisting of 4.4 g L−1 MS with 0.5 g L−1 2-morpholinoethanesulfonic acid (MES), 30 g L−1 sucrose, 0.1 mg L−1 BA, and 8 g L−1 agar.
About 40 ONSEN-like elements were detected in V. angularis and V. radiata (Supplemental Table 1). Of these, three and four copies of ONSEN-like elements were solo-LTRs in V. angularis and V. radiata, respectively. About half of the copies, including these solo-LTRs, were observed to be partial sequences. Even when both the LTRs (partial in some cases) were present, large insertions or deletions were observed, especially in the ONSEN-like elements of V. radiata that possibly lack mobility. Nine copies in V. angularis and six copies in V. radiata were full-length copies without large structural variations, although most of them had stop codons or frame-shift mutations because of small indel variations in their CDS regions. In V. angularis, only three copies of ONSEN-like elements had complete structures, without any disruptive variations (Fig. 1A) although no copy had complete structures in V. radiata. Because accuracies of genome sequences differ among the species, un-sequenced (un-assembled) regions are expected in the whole genome sequences; this was especially observed for V. radiata.
Structure and phylogenetic relationship of ONSEN-like elements in Vigna species. (A) The schematic figure of full-length ONSEN-like elements in V. angularis. GAG: gag polyprotein, PR: protease, INT: integrase, RT: reverse transcriptase, RH: ribonuclease H. probe: A Southern probe designed in the open reading frame (ORF). (B) A phylogenetic tree was made by the neighbor-joining method with synonymous distance estimated by the Nei and Gojobori method. Bootstrap probabilities over 50% are shown along the branches. The copies of V. angularis and V. radiata ONSEN-like elements are indicated by closed and open labels, respectively. The LTR identities are indicated by different label shapes: large circles >99%, small circles ~99–95%, diamonds ~95–90%, squares <90%, and triangles one or no LTRs. Scale bar is shown on left at the bottom of the figure. AtONS copies were included as out-groups. Arrows indicate copies without any disruptive mutations, suggesting that they are active copies.
Furthermore, we analyzed the distance between VaONS and the flanking gene in V. angularis. The average distance was 4187 bp and the distance between the newly inserted VaONS that contained conserved LTRs (p-distance < 0.05) and the flanking gene was 1048 bp (Supplemental Table 1). The distance was shorter than the average distance between the genes in the genome (14 kb: Sakai et al. 2015). Indeed, of the 36 VaONS sequences, seven were within the gene-coding sequences (Supplemental Table 1).
The evolutionary history of ONSEN-like elements in Vigna species is shown in Fig. 1B. For LTR retrotransposons, the sequence divergence within the copies of individual elements between the two LTRs has been proposed as an internal indicator of age, as these are usually identical to each other upon insertion (Kijima and Innan 2010, SanMiguel et al. 1996). Several copies had long external branches suggesting old insertion events. These copies showed small LTR identities or a lack of LTR. Recently-amplified copies (more than 95% LTR identities) and several copies made species-specific clusters. The clusters were supported by more than 80% bootstrap values. The presence of species-specific clusters suggests rapid amplification and degradation of ONSEN-like elements in Vigna species.
To analyze the copy number variations in VaONS, we conducted a Southern blot analysis using the DNA extracted from the accessions of wild and cultivated adzuki beans. About 14 to 16 bands were detected in each accession, of which 12 were common in all the analyzed accessions (Fig. 2A, 2B). The average differences of the number of detected bands between the accessions were 0.886 in the wild accessions and 0.975 in the cultivated accessions. We analyzed the correlation of the band patterns and the origin of the accessions (i.e. Eastern vs. Western and Wild vs. Cultivated). Although there were no significant differences between the accessions from the different geographical origins, there were several significant differences between the wild and cultivated accessions (Table 1). A 9-kb band was specific to the cultivated accessions. Other two bands at 5 kb and 3.3 kb were also represented more in the cultivated accessions.
Variation in the copy numbers of VaONS in wild (A) and cultivated (B) adzuki beans. The number on each of the lanes indicates independent adzuki accessions corresponding to that in Supplemental Tables 2, 3. A gel stained with ethidium bromide (EtBr) is shown at the bottom of each panel as a loading control.
The band a to f correspond to the band in Fig. 2.
+ or − indicate the presence or absence of the listed band in each accessions.
To examine the heat-activation of VaONS, we analyzed the expression of VaONS in two varieties of V. angularis, namely Shumari and Tanba Dainagon. The result showed that VaONS was activated in both the varieties, when subjected to heat stress (Fig. 3A). To detect the presence of an intermediate for transposition of VaONS, we conducted Southern blot analysis with non-digested DNA. The ecDNA was detected in the heat-stressed Tanba Dainagon but not in Shumari (Fig. 3B).
Effect of heat stress on VaONS activation. (A) Heat-induced VaONS transcription. Levels of VaONS transcripts in the two adzuki varieties subjected to 22°C or 40°C were quantified by RT-PCR, with ACTIN2 (ACT2) transcripts used as an internal control. (B, C, D) Southern blot of non-digested DNA loaded to detect extrachromosomal DNA in the two adzuki varieties (B), and in the wild (C) and cultivated (D) adzuki beans in Japan subjected to heat stress. Arrows indicate extrachromosomal DNA (5 kb) of VaONS.
Next, we analyzed the ecDNA in the Japanese accessions of wild and cultivated adzuki beans. The ecDNA was detected in 36 out of the 74 (49%) cultivated accessions and in 12 out of the 38 (32%) wild accessions subjected to heat stress (Fig. 3C, 3D, Supplemental Tables 2, 3). To find a relationship between the accession-specific copy and heat-activation, we compared the presence of polymorphic VaONS bands detected by Southern blot and the synthesized ecDNA (Table 2, Supplemental Tables 2, 3). The result showed that presence of a band specific to the cultivated accessions strongly associated with the ecDNA of VaONS (Fig. 2B, Table 2).
+ or − indicate the presence or absence of the listed band in each accessions.
We also analyzed the geographical distribution of the accessions that synthesized the ecDNA, in Japan. In the wild accessions, ecDNA was detected in four out of nine accessions from East and in eight out of 29 accessions from West (p = 0.29). In the cultivated accessions, ecDNA was detected in 11 out of 15 accessions from East and in 14 out of 33 accessions from West (p = 0.046; Table 3). The result showed that a significant number of cultivated accessions that synthesized ecDNA were collected from East Japan.
+ or − indicate the presence or absence of ecDNA in each accessions.
To analyze the transposition of VaONS, we induced callus from the hypocotyl of heat-stressed Tanba Dainagon, which was obtained from a seed company. The callus was subsequently transferred to an appropriate organ-induction medium and a complete plant was regenerated from it. New VaONS insertions were detected in the plant regenerated from the callus (Fig. 4).
Callus-mediated transposition of VaONS in regenerated adzuki plant. (A) Callus induction and dedifferentiation in Vigna angularis. The days indicate the duration of feeding period after the callus induction. Scale bar indicates 5 mm. (B) Southern blot of VaONS in the regenerated plants. Each lane represents the regenerated individuals derived from two independent plants. Arrowheads indicate new insertions of VaONS. A gel stained with ethidium bromide (EtBr) is shown at the bottom of each panel as a loading control.
On exploring the genome sequences of V. angularis and V. radiata, about 40 regions were found to be ONSEN-related sequences (Supplemental Table 1). Phylogenetic analysis and estimation of the insertion age revealed species-specific clusters, comprising of recently transposed copies of the ONSEN-related sequences. This could have occurred because the amplification rate was high and the amplification was recent in both the species. However, most of the ONSEN-related sequences were observed to be partial fragments and only few regions had the entire sequence of the TE. The genome of V. angularis has three complete copies of ONSEN-related elements, whereas V. radiata has no mobile copies although the presence of active copies is assumed in the non-sequenced genomic regions. The older inserted copies tend to have many disruptive mutations (frame-shifts, novel stop codons, and large insertion–deletions). Loss of active copies could cause the elimination of all the family members because there will be no amplification resulting in new copies. A similar situation was observed in Brassicaceae species (Ito et al. 2013), where Capsella and Cardamine species had no ONSEN-related sequences. This could occur because of the transposition tendencies of the ONSEN family as they transpose close to the genic regions of actively transcribed loci (Ito et al. 2011). Because the ONSEN family sequences transposed to or near the genes in Vigna species (Supplemental Table 1), the transposition could be deleterious and could effectively be silenced or eliminated. The rapid amplification and elimination cycle caused species-specific clustering of relatively young inserted copies of the ONSEN-related sequences (Fig. 1). The shorter external branches in V. angularis, compared to those in V. radiata, also suggest lesser activities in V. radiata as was also suggested in simulation studies (Kijima and Innan 2013, Navarro-Quezada and Schoen 2002).
The copy number of VaONS was conserved in the Japanese accessions indicating that the retrotransposition activity of VaONS was tightly regulated, although it was activated by heat stress. There was no difference of the number of polymorphic bands between East and West varieties although there was significant difference between wild and cultivated accessions. Most polymorphic bands were polymorphic in both the wild and cultivated accessions. Only one and two bands were specific to the wild and cultivated accessions respectively. Because wild adzuki bean is often found around the fields of farmers, outcrossing occurs at around 1% (Yamamoto et al. 2006). Our results also support the ongoing introgression between the wild and cultivated adzuki beans in Japan.
The heat activation of ONSEN was originally observed in A. thaliana and we previously reported the heat-activation of ONSEN-like elements among Brassica species (Ito et al. 2013). In this report, we analyzed the heat-activation of the element in more distant dicot relatives, the adzuki beans. We found that the heat activation was conserved in adzuki species and some ONSEN copies synthesized the ecDNA. As is evident from the sequence analysis, three VaONS copies had complete structure and could be activated by heat stress in adzuki beans. Southern blot analysis showed that more ecDNA was detected in the accessions from East than in those from the West, even among the genetically homogeneous cultivated accessions, suggesting that the gene flow of active copies of VaONS between the cultivated and wild accessions had occurred in adzuki beans (Wang et al. 2004).
The ecDNA was detected in the seedlings subjected to 40°C for 24 h; however, the DNA was not detected in the seedlings subjected to 37°C for 24 h (data not shown). This indicated that the temperature which was sufficient to activate the ONSEN-like elements was variable among the species and VaONS might be less sensitive to heat stress compared to the element in A. thaliana. It is worth mentioning that VaONS was expressed in two varieties of V. angularis, namely Shumari and Tanba Dainagon; however, the ecDNA was detected only in Tanba Dainagon. The ONSEN-like retrotransposons that have stop codons or frame-shift mutations in the CDS regions could be transcribed by heat stress, because the expression of the retrotransposon was regulated by a heat-shock transcription factor that was bound to HRE in the LTR promoter. Our results indicated that the transcriptional activation and the retrotransposition via ecDNA were not always associated.
As is possible for many other TEs (de Araujo et al. 2005, Hirochika 1993, Hirochika et al. 1996, Madsen et al. 2005, Masuta et al. 2017, Planckaert and Walbot 1989, Sato et al. 2011, Yamashita and Tahara 2006, Yilmaz et al. 2014), we could successfully induce retrotransposition of VaONS by tissue culture (Fig. 4). It requires further research to understand the mechanism of the callus-mediated transposition of ONSEN; however, we have reported that the transposition of heat-activated ONSEN during tissue culture was regulated by an alternative mechanism in addition to the RdDM-mediated epigenetic regulation in A. thaliana (Masuta et al. 2017). Although mutagenesis via culture-induced transpositions has problems, including the presence of somaclonal variations in the regenerated plants (Bairu et al. 2011, Larkin and Scowcroft 1981) and requires relatively costly manipulation, it does generate new alleles and would help in expanding the genetic variation of adzuki bean.
To conclude, our results demonstrate that the ONSEN-like elements (VaONS) were present in all the analyzed Japanese accessions of adzuki bean (Vigna angularis), as detected by Southern blot analysis. We found that the heat activation of ONSEN copies was conserved in adzuki species and some ONSEN copies synthesized the ecDNA. Moreover, we could successfully induce retrotransposition of VaONS by tissue culture. Adzuki bean cultivars are always challenged by cold stress, new types of bacterial diseases, and nematode infections. Our findings should shed light on new materials that could be useful in molecular breeding of resistant varieties. Because VaONS is functional in adzuki bean, there might be other active TEs in this and related species. Given the fact that many other species of Vigna are being sequenced (Sakai et al. 2015), it would be worth trying to find such elements for further understanding of genome evolution and for using them as a tool in breeding of new varieties of plants.
We would like to thank Dr. Hitoshi Satou for providing the seeds of Shumari. This work was supported by Suharakinennzaidann, Cooperative Research Grant of the Plant Transgenic Design Initiative, Gene Research Center, the University of Tsukuba, the Joint Research Program of Arid Land Research Center, Tottori University (28C2002), and a Grant-in-Aid for scientific Research on Innovative Areas (JP15H05960) to HI and Grants-in-Aid from MEXT-Supported Program for the Strategic Research Foundation at Private Universities (2015–2019) to AK.