Introduction

Legume plants are a sustainable source of food and feed around the world. They establish beneficial symbiotic interactions with rhizobial bacteria, resulting in special structures, named root nodules. In these structures, rhizobia fix atmospheric nitrogen that plants can use, reducing the need for external fertilizers. Considering the importance of legumes, greater emphasis should be placed on the collection of genetic information for legume species.

Recent advances in “omics” approaches and studies performed on the two model legumes, Medicago truncatula Gaertn. (Cook 1999) and Lotus japonicus L. (Stougaard 2001), are adding to the understanding of legume biology. Both legume models have relatively small genomes, short life cycles, and can be easily regenerated and transformed. These advantages make it possible to analyze the phenotypic and functional features associated with gene and protein up- and down-regulation in intact plants, as well as the expression patterns of promoters and organelle-specific markers in intact legume cells (Boycheva et al. 2015; Revalska et al. 2015).

An important goal of legume genomics is knowledge transfer between model and crop legumes, providing means to improve plant breeding efficiency. Information on certain common traits of legumes, including signal transduction events that lead to nodule formation and synthesis of flavonoids and their glycosides, is considerably transferable from models to crops. Knowledge gained from the model legumes on traits of agronomic importance, such as plant architecture, yield components, and stress tolerance, could have a great impact for the legumes like alfalfa (Medicago sativa) (Zhou et al. 2011), pea (Pisum sativum), and clover (Trifolium sp.). Legume plants and their microsymbionts might even provide a basis for improvement of the performance of other crop species, such as tomato (Solanum lycopersicum), sunflower (Helianthus annuus), cotton (Gossypium hirsutum), corn (Zea mays), and rice (Oryza sativa) (Dakora 2003).

Because of worldwide efforts, the genomes of M. truncatula and L. japonicus are now available (Sato et al. 2008; Young et al. 2011), allowing for comparative genomic analyses. Several technological platforms, such as ethyl methanesulfonate (EMS) mutagenesis combined with targeted induced local lesions in genome (TILLING) techniques (Perry et al. 2003), and microarrays of several thousand non-redundant cDNA clones for transcriptomic and proteomic profiling (Wienkoop and Saalbach 2003), have moved forward functional genomic studies in the model legumes. However, there is still a striking incongruity between the level of advancement of these technologies and the efficiency of exploitation of some basic in vitro procedures for obtaining transgenic plants and their analysis.

Construction and phenotypic screening of mutant libraries is an effective approach to identify the function of unknown genes (Cigan et al. 2005). Different types of mutagenesis, such as T-DNA insertions, chemical mutagens, ionizing radiation, and fast neutron bombardment, have been attempted in legumes; however, with these tools, only a few M. truncatula and L. japonicus mutant lines have been generated and characterized. Imaizumi et al. (2005) demonstrated that T-DNA tagging is possible in L. japonicus but it is labor-intensive because of the relatively low transformation efficiency. Similarly, the maize class II transposon activator and dissociation transposes in the L. japonicus genome (Thykjaer et al. 1995), but the low efficiency of transposition does not permit its utilization as an efficient tag. To date, only one of the reported symbiotic mutants has been isolated by this transposon tagging strategy (Schauser et al. 1999). In contrast, LORE1, an endogenous retrotransposon in L. japonicus, is mobile in regenerated lines and can be used to establish larger mutagenized populations (Fukai et al. 2012). Genome-wide LORE1 mutagenesis and high-throughput insertion detection in L. japonicus have been already reported (Urbanski et al. 2012). Currently, around 200,000 flanking sequence tags (FSTs) have been generated (Lotus base http://users-mb.au.dk/pmgrp), allowing the identification of insertions in the majority of the genes of L. japonicus.

Retrotranspons are a type of mobile elements (class I) that transpose to new sites in the genome via an RNA intermediate. Their primary transcripts are retro-transcribed into DNA copies by a reverse transcriptase (Wessler 2006), and these copies are inserted back into the genome in multiple places (Kumar and Bennetzen 1999). Retrotransposon-induced mutations are relatively stable because there is no excision during replicative transposition. In addition, some retrotransponsons can be highly mutagenic because of prevalent transposition to gene-rich, low-copy regions, making them suitable for tagging genes in plant species with large genomes.

The tobacco (Nicotiana tabacum) retrotransposon Tnt1 is one of the few elements to achieve a saturation mutagenesis in tobacco and heterologous plant genomes. d’Erfurth et al. (2003) first described it during the early steps of in vitro transformation and regeneration in the R108 ecotype of M. truncatula. Later, Tnt1 insertional mutagenesis was successfully used for large-scale mutagenesis of M. truncatula ecotype R108 (Tadege et al. 2008; Cheng et al. 2011), during in vitro regeneration in M. truncatula cv. Jemalong (Iantcheva et al. 2009), and in the important crops lettuce (Lactuca sativa L.; Mazier et al. 2007), soybean (Glycine max [L.] Merr.; Cui et al. 2013), and potato (Solanum tuberosum L.; Duangpan et al. 2013).

The high efficiency of Tnt1 transposition during in vitro regeneration of M. truncatula, resulting in multiple Tnt1 insertions in the newly regenerated lines, encouraged evaluating the Tnt1 tagging as an additional and complementary strategy to the LORE1 system of mutagenesis in the model legume L. japonicus. As a first step, a new protocol for in vitro regeneration and transformation via indirect somatic embryogenesis was developed and used to introduce Tnt1 via Agrobacterium-mediated transformation. To reactivate Tnt1 inserted copies, ‘starter lines’ containing different numbers of inserted Tnt1 and having good regeneration capacity were constructed. Tnt1 transposition frequency was evaluated in the primary regenerated (R0) lines (starter lines), R1 lines, and their T1 progeny. In vitro regeneration resulted in new transposition events in the regenerated R1 lines, and the insertions were stable in T1 progeny. These findings demonstrate that Tnt1 can serve as an additional and alternative to the LORE1 retrotransposon mutagenesis system for the construction of an insertion mutant collection for L. japonicus.

Material and Methods

Plant material and growth conditions

Seeds of L. japonicus ecotype B-129 Gifu, a kind gift of Dr. Hiroshi Kouchi, were scarified with sandpaper and surface sterilized with 70% (v/v) ethanol for 30 s, followed by 0.1% (v/v) mercury chloride (HgCl2) for 8 min, then rinsed at least three times with sterile distilled water, and germinated on Murashige and Skoog (MS) basal medium (Murashige and Skoog 1962) supplemented with 3% sucrose (w/v), 0.7% (w/v) Phyto agar (Duchefa, Haarlem, the Netherlands). Germinated seedlings were then propagated by cuttings. Explants collected from 30–40-d-old in vitro plants were used to initiate callus tissue. In vitro plant material was grown in Magenta™ boxes (60 × 60 × 96 mm; Sigma-Aldrich®, St. Louis, MO) in a growth chamber at 24°C, with a 16-h photoperiod and white cool light, tube intensity of 30 μmol m−2 s−1 (Solar Leading Group Co. Ltd., Jieyang, Guangdong, China).

Composition of plant regeneration and transformation media

Callus induction medium (CIM) and embryo induction medium (EIM) were essentially the same as described by Iantcheva et al. (2009). CIM was based on B5 (Gamborg et al. 1968) solid medium containing 4 mg L−1 2,4-dichlorophenoxyacetic acid (2,4-D), 0.8 mg L−1 kinetin, 1 mg L−1 adenine (6-aminopurine), 500 mg L−1 casein hydrolysate, 500 mg L−1 myo-inositol, 3% sucrose (w/v), and 0.7% (w/v) Phyto agar (Duchefa). EIM was based on MS medium supplemented with 0.9 mg L−1 6-benzylaminopurine (BAP), 0.3 mg L−1 α-naphthaleneacetic acid (NAA), 3% sucrose (w/v), and 0.7% (w/v) Phyto agar (Duchefa). The newly established embryo development medium (EDM) for L. japonicus was based on B5 medium supplemented with 0.2 mg L−1 BAP, 3% sucrose (w/v), and 0.7% (w/v) Phyto agar (Duchefa). Basal MS containing 3% sucrose (w/v), 0.7% (w/v) Phyto agar (Duchefa), and with no plant growth regulators was used for embryo conversion and rooting (ECR) medium.

The pH of all culture media was adjusted to 5.8 with 1 N NaOH before adding the agar. The media were sterilized by autoclaving at 121°C and 103 kPa for 20 min. All cultivations (with an exception of the callus tissue induced in dark) were carried out in a growth chamber at 24°C, with a 16-h photoperiod and white cool light, tube with intensity of 30 μmol m−2 s−1 (Solar Leading Group Co. Ltd.).

Regeneration procedure

Wounded leaves and petioles of 30–40-d-old in vitro plants were transferred onto CIM under light conditions mentioned above or continuous dark. Twenty explants per Petri plate were placed in direct contact with the medium. At day 60, each explant appeared as a ball of friable callus. Embryo induction started after transfer to EIM and lasted about 25–30 d. Development of embryos continued on EDM for two to three passages (each 20 d). Dark green, globular embryos slowly developed and formed cotyledonary leaves after at least two passages (40 d) on this medium. This process continued for 45–60 d. Only embryos with well-developed cotyledonary leaves were transferred to ECR medium for conversion to plantlets. They still looked like a cluster of structures. Subculture (every 25 d) on ECR medium continued until healthy plantlets formed. These plantlets could be separated and maintained on MS basal medium. In total, the optimal period for plant regeneration following this protocol was 150 d.

Bacterial strain and plasmid

For plant transformation, Agrobacterium tumefaciens EHA105 (Hood et al. 1993) was used, carrying the plasmid p1381XC::Tnt1, which is a pCambia1381XC (http://www.cambia.org) vector with the autonomous Tnt1 retrotransposon and a chimeric hptII gene encoding plant resistance to hygromycin. A. tumefaciens EHA105 (pCambia1381XC::Tnt1) was maintained on agar-solidified (1.5%; w/v) YEB nutrient medium with 100 mg L−1 rifampicin and 100 L−1 kanamycin for p1381XC::Tnt1 plasmid selection. Prior to transformation, the cultures were incubated in YEB for 24 h under constant agitation (250 rpm) at 28°C. Bacterial suspensions at an optical density of OD600 = 0.3 were used for inoculation.

Transformation procedure

Leaf and petiole explants detached from in vitro plants were wounded with a scalpel blade and cultured on CIM for 48 h in dark at 24°C before inoculation with bacterial suspension. Explants were inoculated with bacterial suspension for 1 h and co-cultivated for 2 d on solid CIM without selective antibiotics. Then, the explants were transferred to solid CIM medium for callus induction with hygromycin (5.5 mg L−1; HygroGoldTM, Invivogen, San Diego, CA) for selection of the transformed cells and carbenicillin (400 mg L−1; Duchefa) to remove bacterial infection. The selective medium was refreshed every 20 d. Embryogenic calluses were then transferred to EIM and EDM as described above. Hygromycin selection was omitted during the embryo development stage and continued at the same concentration on medium for embryo conversion and rooting. Carbenicillin was reduced to 200 mg L−1 in EIM and EDM and removed from ECR medium. All regeneration and transformation experiments were conducted three times with three replicates.

PCR analysis

Genomic DNA was extracted from transgenic plants using the DNeasy Plant Mini Kit (Qiagen, Germantown), according to the manufacturer’s instructions. PCR amplification to detect the presence of the hptII gene in the genome of the transgenic plants was performed in a reaction containing 1× PCR amplification buffer, 200 μM of each dNTP, 0.5 μM each of forward and reverse primers, 1 U Taq DNA polymerase, and Milli-Q H2O to a final volume of 20 μL. The following amplification conditions were used: [94°C 3 min (1×); 94°C 30 s, 55°C 45 s,72°C 45 s (35×); 72°C 5 min (1×)]. PCR products were evaluated by electrophoresis through 1.0% (w/v) agarose gels. Gene-specific primers for the hptII sequence that were used to amplify a 550-bp product were forward 5′-CCATTCGGACCGCAAGGAATC-3′ and reverse 5′-TCGGCGAGTACTTCTACACAG-3′.

Inverse PCR

Inverse PCR (IPCR) allows amplification of the flanking sequences of Tnt1 copies inserted into the genome (Ratet et al. 2010). Genomic DNA was isolated using the same kit as mentioned above, then digested, self-ligated in a 200-μL reaction volume, and the borders were amplified through a nested (two-step) PCR. Genomic DNA (500 ng) of all tested lines was digested in 50-μL reaction volumes for at least 2 h at 37°C (or overnight at 16°C). The digestion was performed with enzymes EcoRI and MfeI in 1× Tango buffer (Thermo Fisher Scientific, Waltham, MA) and heat inactivated at 65°C for 20 min. Fifty microliter of the digestion reaction were used for ligation (T4 DNA ligase 5U μL−1; Thermo Fisher Scientific) in 200 μL reaction overnight at 16°C. Two nested PCR reactions were performed with primers specific for the borders of Tnt1. IPCR1 [94°C 2 min (1×); 94°C 20 s, 72°C 3 min (30×); 72°C 5 min (1×)] was performed using LTR5 (5′GCCAAAGCTTCACCCTCTAAAGCCT3′) and LTR3 (5′AGTTGCTCCTCTCGGGGTCGTGGTT3′) primers. IPCR2 [94°C 2 min (1×); 94°C 20 s, 60°C 20 s, 72°C 3 min (30×); 72°C 5 min (1×)] was performed using LTR6 5′GCTACCAACCAAACCAAGTCAA3′ and LTR4 (5′TACCGTATCTCGGTGCTACA3′). PCR reactions were done using the Takara Ex Taq™ DNA polymerase enzyme (Lonza Rockland Inc., Rockland, ME). PCR products were analyzed on 1.5% agarose gels.

FST characterization

Cloning and sequencing of the FSTs were performed as described by Ratet et al. (2010).

Materials and reagents

All chemicals and cell culture reagents were purchased from Sigma-Aldrich® (Munich, Germany) unless otherwise specified.

Results

Development of Tnt1 efficient regeneration and transformation protocols for L. japonicus

During regeneration, the presence of a high 2,4-D concentration in the CIM resulted in total dedifferentiation of the plant explants and production of soft and friable callus tissue (Fig. 1A, B ). The use of either light or dark conditions led to the production of massive callus tissue which differed only in color (yellow under dark and greenish under light conditions; Fig. 1A, B ). Approximately 85% of the initial explants (16 of 20 plated) responded positively to the inductive signal (Table 1). Combination of BAP and NAA in the EIM resulted in the formation of embryogenic zones, resembling green spots, which further developed to two to five globular embryos per each callus tissue (Fig. 1C, D ) for a period of 30 d. In the newly established medium for embryo development (ED), the auxin NAA was removed and only BAP triggered further development of globular embryos. Globular embryos slowly developed to torpedos, which formed cotyledonary embryos with emerging cotyledonary leaves after at least two passages (40 d) on this medium. This process continued around 45–60 d. The conversion rate of the globular embryos to torpedo and cotyledonary stage embryos was 50% (27 cotyledonary embryos per 20 explants; Fig. 1E–G ). Only embryos with well-developed cotyledonary leaves either alone or grouped into clusters were transferred to ECR medium for conversion to plantlets. From each 20 initial explants, 14 healthy plantlets were regenerated (Fig. 1H and Table 1). The overall regeneration efficiency of the established protocol was 70%.

Figure 1
figure 1

Basic steps in regeneration of Lotus japonicus from leaf and petiole explants. (A) callus tissue initiated in dark conditions, (B) callus tissue initiated in light conditions, (C) green embryogenic zones on EI medium, (D) cluster of globular embryos on medium for ED, (E) globular and torpedo embryos (arrows, globular; arrowhead, torpedo), (F) cotyledonary embryos with emerging cotyledonary leaves, (G) cluster of cotyledonary embryos, and (H) healthy plantlets.

Full size image
Table 1 Efficiency of the regeneration and transformation procedures for L. japonicus
Full size table

Pretreatment of explants on CIM for 48 h before Agrobacterium infection, using bacterial suspension at OD600 = 0.3, and inoculation for 1 h were found to be optimal conditions for genetic transformation of leaf and petiole explants, preserving their strength and ability for regeneration (data not shown). Selective pressure was kept during the first two regeneration stages—callus and embryo induction, but was removed during embryo development. This led to the successful regeneration of 57 healthy and vigorous, putatively independent transgenic plants, starting from 200 transformed explants. PCR analysis confirmed incorporation of the hptII gene in 86% of the regenerated transgenic lines (Fig. 2A ). The observed 14% rate of escapes (Table 1) could be related to the selection on medium containing 5.5 mg L−1 hygromycin, which did not completely eliminate initial callus induction in control explants but helped identify transformed tissues that grew faster. The transformation step of the protocol resulted in 25% efficiency. In total, 51 starter lines with 1 to 8 copies of Tnt1 were generated.

Figure 2
figure 2

Representative PCR analyses confirming the presence of transgenes. (A) PCR analysis for the presence of hptII gene in the genome of Tnt1 R0 transgenic lines. Lanes 1–23, tested plants; lane 24, negative control (C−); lane 25, positive control (C+); lane 26, DNA marker. (B) Inverse PCR analysis of Tnt1 starter lines (StA-I) with the most common observed profile. Lanes 1–11, Tnt1 positive lines; lanes 4 and 5, St lines 4 and 5 with five and three Tnt1 copies, respectively; lane 12 (−C), negative control; lane 13 (M), DNA marker.

Full size image

Defining the Tnt1 copy number in starter lines of L. japonicas

The Tnt1 copy number for 57 primary transformants (starter lines; St) was analyzed. Most of the transgenic lines (39) showed one band of approximately 600 bp (Fig. 2B ). This fragment corresponds to the flanking sequence of the Tnt1 retrotransposon in the pCambia1381XC:Tnt1 T-DNA that was integrated into the transgenic plants. This further confirmed the transgenic nature of the plants. The fragment of less than 200 bp in St4 and St5 lines (lanes 4 and 5; Fig. 2B ) and St6 and St8 (Fig. 3) corresponded to the two borders of the element plus a few additional base pair and probably represent free (extra-chromosomal) copies of the element present in the tissue of the transgenic plants. The free copies indicate expression and reverse transcription of the retrotransposon in these lines (Fig. S1). In the line St4, three to four additional fragments were detected that corresponded to new insertion sites of the element, indicating that Tnt1 was active in this line during the transformation/regeneration process.

Figure 3
figure 3

IPCR of Tnt1 starter lines. Number of Tnt1 copies: lanes 1 and 2, St1 with one copy; St3 with two copies; lane 3, St6 with eight copies; lane 4, St7 with three copies; lane 5, St8 with four copies; lane 6, St9 with five copies; lane 7, St11 with four copies; lane 8, St12 with seven copies; lane 9, St13 with five copies; lane 11, negative control; lane 12, DNA marker.

Full size image

The results presented in Fig. 3 indicate that some of the regenerated (R0) lines (St7, St8, and St9) did not contain the 600-bp fragment, as expected for the T-DNA construct. However, these lines were PCR-positive for Tnt1 FST. For example, St9 contains multiple FSTs that could represent multiple insertion sites of the Tnt1 element. Among the other lines, some showed only the T-DNA copy (St1) or both the T-DNA copy plus putative extra-chromosomal copies [lines St3, St6, St11, St12, and St13 (Fig. 3) and lines St3, St12, and R6 (Fig. S1)]. This was another indication that transposition had occurred during the transformation process. The characteristics that indicate the presence of T-DNA, the number of Tnt1 inserts, fertility, and marker segregation in the progeny of the R0 and R1 lines, are given in Table 2.

Table 2 Tnt1 transposition characteristics in different regenerated lines and their progeny
Full size table

Tnt1 transposition was reactivated via indirect somatic embryogenesis

To investigate whether tissue culture-induced activation of Tnt1 is applicable for L. japonicus, in vitro regeneration was tested to determine if it could reactivate Tnt1 transposition. All starter lines had conserved the ability to regenerate plantlets via indirect embryogenesis (data not shown). The regeneration protocol described in this work was established particularly for Tnt1 reactivation in L. japonicus. For the Tnt1 reactivation experiments, St3, St6, and St11 were selected (Fig. 3 and Table 2) and their regenerants and T1 progeny were analyzed. The transposition characteristics of the randomly selected regenerants R12, R14, and R18 (derived from St11, with four Tnt1 inserts), and the regenerants R13, R15, and R16 (derived from St3, with two inserts) and R6 (derived from St6, with eight Tnt1 inserts), are presented in Table 2. The regenerated lines R12 and R14 contained three and four new Tnt1 inserts when compared to St11. In the regenerated line R13, two new inserts were detected, and the line R15 four new transposition events were detected. The new transposition events, resulting from Tnt1 reactivation in the L. japonicus ecotype B-129 Gifu background, varied from 1 to 4, significantly less than reported for M. truncatula cv. Jemalong, where 8 to 12 new transpositions occurred in tissue culture experiments using a starter line with 8 copies.

Tnt1 inserts were heritable in T 1 progeny

Vigorous in vitro plants from starter and regenerated lines were transferred to greenhouse in order to obtain their offspring. The plants had regular flowering time and pod production. The only exceptions were for some of the regenerants (line R6) deriving from St6 that were sterile. Self-fertilized T1 progeny were obtained from all other starter and regenerated lines with new transposition events.

Transgenic T1 seeds from several starter lines and seeds of regenerants were grown on selective ECR medium with 5.5 mg L−1 hygromicin. All tested lines segregated in the ratio 3:1 with an exception of T1 progeny of lines St6 and St13, which segregated in a 1:1 ratio. This deviation could be explained by the small population of analyzed seeds (only 12) or as a result of chromosomal rearrangements.

IPCR was utilized to confirm that the transposition profile of some of the starter lines and showed that their progeny did not change during plant growth in a greenhouse. The transposition profiles of T1 regenerants R14 and R12 and starter lines St3, St11, St8, and St13 are shown in Fig. 4. There were no new transposition events during plant growth and in the obtained progeny, suggesting that the Tnt1 element is not active during normal plant life cycle.

Figure 4
figure 4

IPCR of selected starter lines, regenerated lines and T1 progeny: lane 1, St3 in greenhouse; lane 2, T1 of St3; lane 3, St11 in greenhouse; lane 4, T1 of St11; lane 5, T1 of R12; lane 6, T1 of R14; lane 7, T1 of St8; lane 8, T1 of St13; lane 9, R6 regenerant of St6; lane 10, negative control (C−); lane 11, DNA marker.

Full size image

Sequencing of FSTs

To confirm the nature of the FSTs deduced by IPCR, FSTs were cloned and sequenced for lines R6, T1 St3, and T1 St12 (Fig. S1). Six fragments from each line were sequenced. Ten fragments corresponding to free copies of the Tnt1 element with only a few base pairs between the two borders were obtained. Similarly, four fragments corresponding to the Tnt1 junction with the T-DNA vector were recovered. From line T1 St3, four sequences were obtained that represented the same FST fragment and which showed Tnt1 insertion in a putative gene. This sequence (493 bp) is not available in the Lotus genome present in public databases. It corresponds to a putative gene coding for a cytochrome P450, highly homologous (90% amino acid identity) to the cytochrome P450 (XP_006596023.1) gene from Glycine max. The sequence of this IPCR fragment shows the 5-bp duplication (TTTTC) characteristic of Tnt1 transposition. This result clearly shows that Tnt1 was active in this line and inserted into a gene.

Discussion

Building large mutant collections for different plant species contributes to identification of mutated genes and an understanding of gene function. During the last two decades, different tools have been developed for efficient generation of mutant collections. Insertional mutagenesis is one of the most powerful tools to generate mutants and understand gene function in plants. T-DNA has been successfully applied as a mutagen in Arabidopsis thaliana (Alonso et al. 2003), but for plants with larger genomes, like model legumes, it is not practical to use this approach.

The use of transposable elements is a good alternative strategy for generation of large insertional mutant collections and has become a powerful tool for gene tagging in model legumes and crops (Mathieu et al. 2009; Cheng et al. 2011; Cui et al. 2013; Grandbastein 2015). Activity of Tnt1 and Tto1 retrotransposons has been detected in interspecific hybrids of tuber-bearing species of Solanum, and represents a source of genetic variability, important for adaptation of potato species (Paz et al. 2015).

In the last decade, Tnt1 has been successfully used as a valuable tool for building M. truncatula mutant collections and applied in forward and reverse genetic studies (Tadege et al. 2008; Zhou et al. 2011; Bourcy et al. 2013; Veerappan et al. 2016). Phenotypic screening of the Tnt1 mutant lines is an efficient way to explore functions of new genes or new regulatory pathways. Analysis of morphological characteristics of M. truncatula mutants generated by Tnt1 retrotransposon insertions has shown that leaf phenotypic characteristics are influenced by the Tnt1 insertions and display visually recognizable phenotypes (Vassileva et al. 2010).

In this paper, Tnt1 insertional mutagenesis has been shown capable of producing collections of insertional mutants in L. japonicus. Application of Tnt1 mutagenesis in the second model legume L. japonicus could provide additional tagged insertion mutant populations and serve as an alternative of the existing LORE1 strategy. The constructed Tnt1 mutant lines will further enrich the legume reverse genetic resources. These mutants may be a valuable resource for functional and comparative genomic studies in other legumes.

The obtained regeneration/transformation protocol was mainly developed to enable efficient Tnt1 incorporation and transposition into the L. japonicus genome. The current study confirmed the importance of the synthetic auxin 2,4-D and the callus induction step for successful Tnt1 transposition, which was explored during the construction of Tnt1 mutant collection for M. truncatula cv Jemalong. Under these optimized conditions, the frequency of transposition of the Tnt1 retrotransposon in the M. truncatula genome was 80–85% (Iantcheva et al. 2009). The present study showed that the initiation of soft and friable callus in the presence of a high concentration of 2,4-D (4 mg L−1) was an important prerequisite for complete dedifferentiation of starting explants (leaf and petiole) of the model legume L. japonicus. Other works related to regeneration and genetic transformation of L. japonicus confirm the importance of the plant hormones NAA and BAP and the hypocotyl as a starting explant or a combination of auxin and cytokinins for dedifferentiation and production of poor callus from root explants (Lohar et al. 2001; Lombari et al. 2003). The published protocols point to the induction of compact callus tissue followed by organogenesis and shoot regeneration.

The described regeneration and transformation protocols via indirect somatic embryogenesis had relatively high transformation efficiency (25%) and allowed producing 51 independent Tnt1 starter lines. The number of Tnt1 copies in these starter lines (1–8) was much smaller than reported for M. truncatula (up to 30), for soybean (4–19), and for potato (1–20) (d’Erfurth et al. 2003; Cui et al. 2013; Duangpan et al. 2013). Three of the starter lines were PCR positive for Tnt1 FST, suggesting that Tnt1 inserted into the L. japonicus genome independently from T-DNA. This result is similar to previous reports for M. truncatula and potato (d’Erfurth et al. 2003; Duangpan et al. 2013).

The Tnt1 retrotransposon could be a useful tool for building a collection of L. japonicus mutant lines only if Tnt1 insertional activity remains inactive during plant growth and if new transpositions do not occur in subsequent generations. The first evidence of in vitro regeneration of transgenic lines containing the Tnt1 retrotranspososn was shown to be sufficient to activate Tnt1 transposition in M. truncatula ecotype R108 and for cultivar Jemalong (d’Erfurth et al. 2003; Iantcheva et al. 2009). In vitro regeneration-induced activation of Tnt1 has been reported also for other plant species (Courtial et al. 2001; Mazier et al. 2007; Cui et al. 2013; Duangpan et al. 2013). This approach is very powerful for a large-scale mutagenesis since in vitro regeneration is faster and more efficient compared to genetic transformation. The regeneration protocol for L. japonicus developed in this study induced the transposition activity of Tnt1 and the insertion of the element in different locations in the Lotus genome and showed activation of Tnt1 via in vitro regeneration. Activation of different tobacco retrotransposons was shown to be specifically modulated by the elicitor cryptogein (Anca et al. 2014) and enhanced by osmotic pretreatment of the starting explants (Iantcheva et al. 2009).

The number of new transpositions observed in this study (two to four) is comparable to the efficiency of the endogenous retrotransposon LORE1 (Fukai et al. 2012), where the average number of the new insertions per plant was 2.6. Despite the low transposition frequency, tissue culture-induced activation of Tnt1 after further optimization and selection of lines devoid of background mutations could offer a promising basis for the construction of large mutant populations of L. japonicus.

Conclusion

The present study showed incorporation of the retrotransposon into L. japonicus ecotype B-129 Gifu and that starter lines with different copy numbers of Tnt1 were obtained. In vitro regeneration from starter lines containing from two to eight copies of Tnt1 resulted in new transposition events, increasing the number of insertions per line. The inserted Tnt1 copies remained inactive during normal plant life cycle and were stably inherited in the progeny. These findings can serve as a methodological basis for the construction of a large insertional mutant collection for L. japonicus. The efficient transformation and regeneration protocols developed for this study can be used by researchers producing transgenic legume lines.