Efficient transposition of the Tnt1 tobacco retrotransposon in the model legume Medicago truncatula

Authors


For correspondence (fax +33 169 82 36 95; e-mail Pascal.Ratet@isv.cnrs-gif.fr).

Summary

The tobacco element, Tnt1, is one of the few active retrotransposons in plants. Its transposition is activated during protoplast culture in tobacco and tissue culture in the heterologous host Arabidopsis thaliana. Here, we report its transposition in the R108 line of Medicago truncatula during the early steps of the in vitro transformation–regeneration process. Two hundred and twenty-five primary transformants containing Tnt1 were obtained. Among them, 11.2% contained only transposed copies of the element, indicating that Tnt1 transposed very early and efficiently during the in vitro transformation process, possibly even before the T-DNA integration. The average number of insertions per transgenic line was estimated to be about 15. These insertions were stable in the progeny and could be separated by segregation. Inspection of the sequences flanking the insertion sites revealed that Tnt1 had no insertion site specificity and often inserted in genes (one out of three insertions). Thus, our work demonstrates the functioning of an efficient transposable element in leguminous plants. These results indicate that Tnt1 can be used as a powerful tool for insertion mutagenesis in M. truncatula.

Introduction

Leguminous plants represent an important source of protein for human and animal nutrition and combined nitrogenous compounds for the biosphere (Vance, 2001). The use of leguminous model plants (Medicago truncatula and Lotus japonicus) has allowed the recent identification of several genes required for symbiosis with bacteria and fungi (Endre et al., 2002; Krusell et al., 2002; Nishimura et al., 2002; Schauser et al., 1999; Stracke et al., 2002). Insertion mutagenesis will help to identify further symbiotic genes as well as genes of agronomic importance.

Large-scale transformation by vacuum infiltration is not yet feasible in Medicago truncatula; therefore, T-DNA mutagenesis is not suitable for the construction of large mutant collections. Transposable elements represent an attractive alternative for constructing such mutant collections (Sundaresan, 1996) because the number of transgenic lines to be produced is not a limiting factor when using these mobile elements.

Two classes of transposable elements have been used for insertion mutagenesis. Class I elements, also called retrotransposons, transpose through an RNA intermediate, which is reverse-transcribed into a linear double-stranded DNA before its integration into the host genome (copy and paste mechanism). Class II elements, also called DNA transposons, transpose through a DNA intermediate (cut and paste mechanism).

Transposon mutagenesis using DNA transposon has multiple advantages. The lines containing the active mobile element can be multiplied to prepare mutated populations on a large scale. In addition, it is possible to demonstrate genetic linkage between the inserted transposon and the mutant phenotype by restoration of the wild-type phenotype after excision of the transposable element. However, DNA transposons generally transpose near their initial integration site so that the mutagenesis of the genome is not entirely random, unless starting with independent plants carrying transposons located at different sites of the genome. Using the maize transposable element Ac/Ds as a mutagen, Schauser et al. (1999) have described the only legume symbiotic mutant obtained so far by insertion mutagenesis.

More recently, several retrotransposons (class I elements) were used as mutagens in plants. Retrotransposons are divided into two subclasses: LTR retrotransposons carrying long-terminal repeats (two direct repeats at both ends) and non-LTR retrotransposons without these repeats. LTR retrotransposons are related to retroviruses as they share the same DNA structure and their life cycles are very similar. Autonomous LTR retrotransposons encode functions required for their own replication and transposition, the capsid protein (gag region), a protease (PR), an integrase (IN) and the reverse transcriptase-RNaseH (RT). As members of this class of elements transpose via a copy and paste mechanism, they are able to invade different host genomes. For example, at least 70% of the maize genome is composed of LTR retrotransposons (Bennetzen, 2000). Transposition of some of these elements is activated by biotic and abiotic stresses, and interestingly by tissue culture. Therefore, different retrotransposons were used for large-scale insertion mutagenesis in different model plants, using their capacity to transpose during tissue culture. For example, Tos17, an endogenous retrotransposon of rice, and Tnt1 and TtoI, two retrotransposons of tobacco, were used for gene tagging in rice and Arabidopsis thaliana, respectively (Courtial et al., 2001; Okamoto and Hirochika, 2000; Yamazaki et al., 2001). These studies demonstrated that retrotransposons efficiently transpose into genes of the heterologous hosts and that the target-site sequences exhibit moderate or no consensus. In addition, they do not transpose in the vicinity of their original location, but are rather dispersed in the genome as a result of their mode of transposition requiring a cytoplasmic phase. Thus, these elements seem to be good candidates for gene tagging in leguminous plants.

Using a very efficient transformation–regeneration protocol for M. truncatula (Trinh et al., 1998), we initiated insertion mutagenesis in the R108 line of M. truncatula using the tobacco LTR retrotransposon Tnt1. Tnt1 was first isolated after its transposition into the NiaD gene of tobacco (Grandbastien et al., 1989). It is an autonomous 5.3 kb long copia-like LTR element that creates a 5 bp duplication upon insertion. Tnt1 transposition was induced in cultures of tobacco protoplast (Pouteau et al., 1991) and during in vitro transformation of A. thaliana (Courtial et al., 2001). Its integration seems to be random as no obvious site specificity was observed and genes were frequently hit. The insertions are stable and most of the time unlinked as a result of the replication cycle of LTR retrotransposons (Courtial et al., 2001). All these characteristics make Tnt1 a good candidate for insertion mutagenesis in other plants.

Here, we report the behaviour of Tnt1 in M. truncatula and demonstrate that Tnt1 transposes actively during in vitro transformation, generating between 4 and 30 insertions per plant. These insertions are stable during the life cycle of M. truncatula. Most of them are genetically independent and can be separated by recombination. We also show that Tnt1 seems to transpose preferentially in genes, and that Tnt1 multiplication by transposition can be re-induced by tissue culture, making this element an attractive candidate to set up a large-scale insertion mutagenesis in M. truncatula.

Results

Tnt1 transposes early during in vitro transformation of M. truncatula

Previous studies in tobacco and Arabidopsis have demonstrated that Tnt1 transposes during tissue culture (Courtial et al., 2001; Melayah et al., 2001). To investigate whether this transposable element could also transpose in M. truncatula, we transformed the R108 line by using the in vitro transformation–regeneration method developed in our laboratory (Trinh et al., 1998). To accomplish this, leaf explants were transformed using the EHA105 Agrobacterium tumefaciens strain (Hood et al., 1993) carrying the Tnk23 plasmid, a derivative of the pBin19 binary vector containing an autonomous copy of Tnt1 (Lucas et al., 1995). By selecting for kanamycin resistance (40 µg ml−1), 537 plants were regenerated and named as tnk lines. Because we previously observed that a large percentage of the regenerants escaped the selection when we used a T-DNA conferring resistance to kanamycin in Medicago (data not shown), each regenerated plant was tested for the presence of the nptII gene (conferring kanamycin resistance) and the Tnt1 retrotransposon by PCR, using the oligonucleotide pairs Km5/Km3 and ltr1/ltr2 (Figure 1 and Experimental procedures). Indeed, among the regenerated plants tested (502 out of 537), only 46.7% amplified the expected fragments with at least one of the two couples. Among them, 4.2% were positive only for the nptII gene and thus probably contained a truncated version of the T-DNA (Tinland, 1996). Of the regenerated plants, 31.3% were positive with the two couples of primers, indicating that they had probably integrated a complete T-DNA. Surprisingly, 11.2% of the regenerated plants were only positive for Tnt1, indicating that they carried only the retrotransposon. To confirm these results, plants from each class were analysed by Southern blot experiment (Figure 1). The genomic DNAs of six independent lines positive only for Tnt1 by PCR, one line positive only for the nptII gene, one line containing an intact copy of the T-DNA and the wild-type line R108 were digested with the DraI restriction enzyme and analysed by Southern blot, using an internal fragment of Tnt1 (tntA-B) and a fragment of the nptII gene (Km3-Km5) successively as probes. The lines positive only for Tnt1 by PCR exhibited the expected 2218 bp internal fragment of Tnt1 but did not hybridize with the nptII probe. Similarly, the line positive only for nptII by PCR hybridized only with the nptII probe, indicating that in this type of plant a truncated fragment of the T-DNA construct corresponding to the nptII gene was transferred to the plant genome. These data indicate that Tnt1 transposed very early during the in vitro transformation procedure, independently of the integration of T-DNA into the genome. Thus, this result indicated that Tnt1 was active for transposition in the M. truncatula tissue culture. In addition, the differences in intensity observed for the hybridizing internal fragments of Tnt1 indicated that these plants contained variable copy numbers of Tnt1. For instance, lines tnk85, tnk87 and tnk114 probably contained multiple copies of the retrotransposon (Figure 1).

Figure 1.

Southern blot analysis of tnk primary transformant lines obtained with the tnk23 binary vector.

Between 5 and 10 µg of total DNA from different tnk T0 lines was digested with the DraI restriction enzyme and analysed by Southern blot.

(a) Structure of the tnk23 T-DNA introduced in Medicago truncatula. The grey arrows represent the Tnt1 long terminal repeats (LTRs). The tntA-B and nptII probes are represented by black lines under the T-DNA scheme. D: DraI restriction sites. Primers Km5/Km3 and ltr1/ltr2 are indicated by arrows under the T-DNA scheme. RB: right border, LB: left border.

(b) Hybridization with the nptII probe. The size of the hybridizing fragments indicates that the DraI restriction site present at the RB was lost upon T-DNA insertion.

(c) Hybridization with the tntA-B internal fragment probe.

(d) PCR amplification results obtained from each plant with the ltr1/ltr2 and the Km5/Km3 primer pairs. MW: molecular weight marker. The size of the molecular weight fragments is indicated in kilo bases (kb) on the right side of the figure.

To investigate the copy number of Tnt1 in the regenerated plants, 41 T0 lines were analysed by Southern blot experiment. Genomic DNA was digested by the restriction enzyme NcoI cutting at position 1488 of Tnt1 once and hybridized with the Tnt1 C-D probe (Figure 2a) corresponding to the 3′ internal part of the element (without the LTR region). For all the plants tested, we observed the presence of multiple bands corresponding to independent Tnt1 insertions in the genome (Figure 2b). In plants carrying an intact T-DNA copy, we generally observed a 5424 bp large band of stronger intensity corresponding to the Tnt1 copy present in the T-DNA. This result indicated that this T-DNA construct frequently inserted in the M. truncatula genome in multiple copies, as previously reported for the pBin19 derivatives (Scholte et al., 2002). This analysis also revealed that the number of Tnt1 insertions independent from the T-DNA varied between 4 and more than 30 (Figure 2; data not shown) per plant, indicating that the transposable element had efficiently transposed during the transformation procedure. Generally, the lines carrying the T-DNA had more Tnt1 copies than those containing only Tnt1 (Figure 2b). In all the lines carrying T-DNA, extra copies of the Tnt1 element were observed, reflecting the high transposition efficiency of the element during tissue culture.

Figure 2.

Southern blot analysis of tnk primary transformants containing Tnt1.

(a) Structure of the tnk23 T-DNA introduced in Medicago truncatula. The tntC-D and tntG-H probes are represented by black lines under the T-DNA. N: NcoI restriction sites. The size of the molecular weight fragments is given on the right of the figure.

(b) Total DNA of eight independent T0 transgenic plants was digested with NcoI and hybridized with the tntC-D probe.

The position of the Tnt1 copy carried by the T-DNA is indicated by a black star. The position of Tnt1 linear intermediates of replication is indicated by a black arrow.

In regenerated M. truncatula plants, Tnt1 is expressed but no longer transposed

To use Tnt1 for insertion mutagenesis in M. truncatula, it is necessary that this element is not amplified to a very high copy number in the T0 plants as well as in their progeny (T1 and T2). Its activity (transcription, replication and transposition) was thus analysed in the transgenic plants.

Transcription of Tnt1 in leaves, flowers and roots was analysed by Northern blot experiment. As shown in Figure 3, Tnt1 was expressed in these different tissues, albeit at different levels. Transcripts were much more abundant in roots and old leaves than in young leaves and flowers. Transcripts of similar size were detected in old leaves of M. truncatula and in leaves of A. thaliana induced by CuCl2 for Tnt1 expression (Mhiri et al., 1997). The most abundant one (Figure 3) corresponds to the full-length Tnt1 mRNA. The other transcripts probably correspond to truncated versions of this mRNA. These results indicate that the pattern of expression of Tnt1 is similar in M. truncatula and A. thaliana (Figure 3a). Comparison of the different transgenic lines of M. truncatula indicated that the level of expression varied, with some lines strongly expressing Tnt1 while others expressed it at a low level. These different levels of transcript accumulation did not correlate with the number of Tnt1 copies in each line (data not shown) and might have resulted from position effects of the transgenes and/or from the presence of modified versions of Tnt1. This pattern of expression of Tnt1 was maintained in the progeny of the transgenic lines (Figure 3b,c).

Figure 3.

Analysis of the expression of Tnt1 in the T0 and T1 lines by Northern blot.

For each sample, 5 µg of total RNA was loaded on the gel and the blot was hybridized with the tntC-D probe (Figure 2).

(a) Tnt1 expression in a transgenic Arabidopsis thaliana line induced by CuCl2 (first lane) and RNA from old leaves of different tnk23 T0 lines of Medicago truncatula.

(b) Tnt1 expression in young leaves (YL), old leaves (OL) and flower (Fl) of four different T0 transgenic lines.

(c) Tnt1 expression in young leaves (YL), old leaves (OL) and roots (R) of four different T1 plants originating from three independent transgenic lines. Note the difference of expression between different organs. Blot C was exposed for a shorter time than blots A and B. The 4.7 kb transcript of Tnt1 is indicated on each blot.

The replication of Tnt1 in the transgenic plants was assayed by Southern blot and inverse PCR (I-PCR) experiments. As observed previously in A. thaliana (Courtial et al., 2001; Feuerbach et al., 1997), the Southern blot experiment revealed that linear reverse transcription products of Tnt1 were detectable only in the DNA sample of T0 plants corresponding to old leaves (data not shown). In the samples isolated from younger leaf tissue of T1 plants, these linear DNA molecules were generally not detectable. Thus, the presence of Tnt1 reverse transcription products correlates with the transcription of the element. I-PCR experiments (see below) also allowed us to detect 2-LTR circles (data not shown).

In Arabidopsis, the presence of these replicating forms in the vegetative tissues did not result in the integration of new germinal Tnt1 copies in the progenies of the transformed plants (Courtial et al., 2001). To check whether this behaviour of Tnt1 was the same in M. truncatula, we searched for germinal transpositions in the progeny of different lines. Southern blot experiments did not reveal new insertions in a total of 56 progeny plants originating from four transgenic lines (33 T1 plants from line tnk66, 10 T1 plants from line tnk88, 7 T1 plants from line 171, and 6 T1 plants from line tnk23), indicating that Tnt1 did not transpose at a detectable frequency in the germ cells of M. truncatula (data not shown). Similarly, we could not detect somatic transposition by Southern analysis in the root DNA material despite the high level of expression of Tnt1 in this organ (data not shown).

Tnt1 efficiently transposes into coding regions

Genes in M. truncatula are probably clustered in the euchromatic regions of the genome, representing only 25% of the complete genome (Kulikova et al., 2001), suggesting that the transcribed regions probably represent less than 15% of M. truncatula genome. Thus, Tnt1 can be used as an efficient mutagen if it gets inserted in genes rather than in non-coding regions of the genome. To test its mutagenic activity, we isolated 38 different Tnt1 insertion sites from 14 independent transgenic lines (Table 1) using PCR-walk and I-PCR techniques. As we used only EcoRV restriction digest for PCR-walk and a combination of the EcoRI and MfeI restriction enzymes for I-PCR, we were not able to isolate all the insertion sites for each line. However, among these 38 sequences, 12 (30%) showed high similarities with M. truncatula ESTs and/or genes of other plant species (Table 1). If Tnt1 insertion in the M. truncatula genome had occurred at random, the tagging efficiency should have been 15% instead of 30%, suggesting that Tnt1 preferentially inserts in transcribed regions.

Figure 1.

Characterization of Tnt1 integration sites by I-PCR and PCR-walk

The I-PCR technique allowed us to isolate the two sides of the insertion sites. In this way, we could detect the typical 5 bp duplications characteristic of the Tnt1 insertion mechanism and thus confirm that the transposition mechanism is conserved and functional in M. truncatula. Moreover, the sequences of the 5 bp duplications in the 38 isolated integration sites were all different except two (Table 1), indicating that there is no sequence specificity of insertion for Tnt1 in M. truncatula although the element has preference for AT-rich regions.

Tnt1 insertions are genetically independent and follow Mendelian segregation

Genetic independence of the different insertions in one transgenic line would indicate that no hot spot regions for Tnt1 insertion exist in the M. truncatula genome, and that crossing to wild-type plants will allow the isolation of insertion responsible for the mutation of interest in future. Thus, to find out whether the different Tnt1 insertions segregate in the progeny of different transgenic lines, we analysed their presence in the T1 plants by Southern blot or PCR experiments.

Selfed progenies of four lines analysed by Southern blot hybridization indicated that the Tnt1 insertions segregated genetically (data not shown). However, as the Southern blot experiment was not accurate enough to separate all the different insertions present in one transgenic line, the PCR technique was used to follow the segregation of Tnt1 element at the loci isolated in the previous experiment. In addition, by using this technique, more plants from the progeny could be analysed. Based on the sequence obtained by I-PCR from some of the tagged loci, different primer pairs were designed in order to amplify four different insertion sites in line tnk176 and six others in line tnk48 (Figure 4). Southern blots of these two lines indicated that they contain more than 20 Tnt1 insertions (data not shown). For each locus tested, except one, by using two oligonucleotide pairs (one specific of the Tnt1 insertion analysed and one for the wild-type locus), plants with a wild-type locus, plants heterozygous for the insertion and homozygous mutant plants were observed. According to the chi-square test, the segregation pattern of nine out of 10 insertions tested corresponded to a ratio of 1 : 2 : 1 (1/4 wild type, 1/2 heterozygous and 1/4 homozygous for the insertion; Figure 4) as expected for a locus with two alleles. Insertion tnk48-5 exhibited a ratio of 1 : 2 (1/3 wild type and 2/3 heterozygous), suggesting that the mutation is embryo lethal, as no homozygous mutant locus was detected among the 45 plants tested. Interestingly, the wild-type sequence surrounding this insertion corresponds to a M. truncatula EST expressed only in the reproductive tissue (EST BI310111; Table 1).

Figure 4.

Tnt1 segregation in lines tnk48 and tnk176.

The segregation of different Tnt1 insertions in the progeny of the two lines was analysed by PCR. Grey squares correspond to wild-type homozygous plants, black squares correspond to plants homozygous for a given Tnt1 insertion and white squares correspond to heterozygous plants. The black and the grey stars indicate the genetically linked insertions. The white star indicates lethal insertion. For each insertion, the observed segregation data as well as the theoretical values (th) expected for a 1 : 2 : 1 segregation are given at the bottom of the figure.

By applying the chi-square test, we demonstrated that in these two transgenic lines (tnk48 and tnk176), most of the insertions tested were genetically independent from each other. However, in the line tnk48, the insertions tnk48-1 and tnk48-12 were genetically linked in cis and the genetic distance was 0.1 UR (Figure 4). The chi-square analysis for genetic linkage of tnk48-5 and tnk48-2 insertions was incompatible with both hypotheses: linkage and absence of linkage. This result may be the consequence of the small size of the analysed sample. However, these insertions seem to be genetically linked in trans, as we observed an excess of homozygous plants for the insertion tnk48-2 relative to the wild-type locus corresponding to the insertion tnk48-5 (Figure 4).

For all the insertions analysed, recombination was observed indicating that for each pair of insertions tested, at least one recombination event could be detected (Figure 4). Thus, these results indicate that generally Tnt1 insertions segregate independently, and are mostly distributed in the gene-rich regions of the different chromosomes of M. truncatula. In addition, this shows that for a given line with an insertion of interest, it would be possible to significantly reduce the number of Tnt1 copies by crossing to the wild type.

Tnt1 transposition can be re-activated by in vitro culture

To test whether increasing the size of the tagged population would be possible by activation of Tnt1 copies already present in the M. truncatula genome, we re-submitted leaf explants of lines tnk66 and tnk88 to the process of in vitro regeneration. Three regeneration conditions were used in order to know which step of our transformation protocol is responsible for triggering Tnt1 transposition. In the first condition (RI), sterilized leaves were cut and directly placed on SHMab (callus inducing solid medium, Trinh et al., 1998). In the second condition (RII), sterilized leaves were cut and vacuum-infiltrated in SHMab liquid medium and then placed on SHMab solid medium. In the third condition (RIII), sterilized leaves were cut, vacuum-infiltrated in SHMab liquid medium containing the A. tumefaciens strain EHA105 and then placed on SHMab solid medium. Plants were then regenerated on SHM2 medium (Trinh et al., 1998) and transferred to the greenhouse. Several T1 plants from lines tnk88 and tnk66 were used for the experiments. The line tnk88 contains few copies of Tnt1 as well as a complete T-DNA copy. The line tnk66 contains multiple copies of Tnt1 but no T-DNA. The genomic DNA of individual plants regenerated from these two lines under the three in vitro culture conditions was digested by the restriction enzyme NcoI and analysed by Southern blot using the tntG-H probe (Figure 2; Experimental procedures). Because Tnt1 had already segregated in the different T1 plants used for the three assays, we pooled their genomic DNA and used it as a control (pool T1). New insertions were obtained for the two lines under the three tested conditions. The results obtained for line tnk88 under conditions RI and RIII are shown in Figure 5. The presence of new insertions in the regenerated plants (lanes 2–3 and 5–6 in Figure 5) as compared to the pooled T1 DNA sample (lane 1 in Figure 5) showed that Tnt1 transposed again, and that tissue culture conditions (RI condition) are sufficient for the activation of the element's transposition.

Figure 5.

Re-mobilization of Tnt1 transposition by a second step of in vitro culture.

Genomic DNA was digested by NcoI restriction enzyme and hybridized with the tntG-H probe (Figure 2). The first lane (tnk88 pool T1) corresponds to a pool of the genomic DNA of the different tnk88 T1 plants used for the in vitro regeneration procedure. Tnk88-2 to tnk88-5 represent independent plants obtained after in vitro regeneration using two different conditions (see text). MW: molecular weight marker lane; R108: lane corresponding to the wild-type line R108; RI: regeneration condition I; RIII: regeneration condition III.

The tnk66 line originally contained transposed copies of Tnt1 but no T-DNA. The presence of new transposition events in this line (data not shown) indicates that during this second tissue culture step, at least one of the transposed copies was still active for transposition.

Tnt1 is expressed in M. truncatula Jemalong line J5

The M. truncatula line Jemalong is another ecotype that has been used to obtain symbiotic mutants induced by physical and chemical agents (Catoira et al., 2001; Penmetsa and Cook, 1997) as well as for most of the molecular studies on this plant. However, so far the transformation–regeneration process of this line has been rather inefficient and time consuming (Kamatéet al., 2000).

In order to rapidly test the transposition activity of Tnt1 in this line, we transformed leaf and flower explants of R108 and J5 lines using the A. tumefaciens strain EHA105 carrying the pHLV5501 plasmid (Courtial et al., 2001). This plasmid contains a translational fusion between Tnt1 and the gus-intron reporter gene. This construct allowed us to follow the transcriptional activity of the retrotransposon during the early steps of the in vitro culture. Plasmids pMS1 and p35S-gusI (Experimental procedures) were used as negative and positive controls, respectively. The expression of the LTR-gus fusion was followed, in three independent experiments, for 15 days after the beginning of the transformation in leaf (day 5 and day 15, shown in Figure 6) and flower (data not shown) explants of the two lines. Leaf explants were used in the above study to produce the transgenic R108 plants, but as we have been using flower explants for transformation–regeneration of the Jemalong line (Kamatéet al., 2000), these were also tested. According to the X-gluc staining observed in the leaf and flower explants (Figure 6 and data not shown), we concluded that the time course of the expression pattern of the Tnt1 promoter was the same in both lines. The expression started 2 days after infiltration of the explants as it was reported for Arabidopsis (Courtial et al., 2001) and was still visible after 15 days. These results indicated that Tnt1 is also transcriptionally active in Jemalong. The expression of Tnt1 was stronger in the leaf explants of J5 but was fairly similar in the flower explants of the two ecotypes (data not shown). Introduction of Tnt1 in this line, already initiated in the laboratory, will thus allow large-scale insertion mutagenesis in it and will facilitate allelism tests with the already existing mutants.

Figure 6.

Histochemical GUS staining of leaf explants of R108 and J5 lines at different times after in vitro transformation using plasmids pHLV5501, pMS1 or p35S-gusI.

(a) Explants stained for GUS activity 5 days after Agrobacterium infiltration.

(b) Explants stained for GUS activity 15 days after Agrobacterium infiltration.

Discussion

Behaviour of Tnt1 in M. truncatula

We demonstrated that the retrotransposon Tnt1 transposes actively in M. truncatula during the early steps of the in vitro culture, as in A. thaliana (Courtial et al., 2001). The regeneration of plants containing transposed Tnt1 copies, but no T-DNA sequences, indicates that Tnt1 can transpose during the very early steps of the in vitro culture, and that this transposition is independent of the T-DNA integration in the genome, suggesting that Tnt1 was expressed just after the infection of the plant cell by the single-stranded form of the T-DNA. The relatively high percentage (11.2%/0.645 = 17.4%) of plants containing Tnt1 among the escaped plants indicates that this element is present in 17% of the regenerants obtained without selection. This indicates that, in principle, Tnt1 can be used as a transformation vector without selection.

The number of Tnt1 insertions varies between four and more than 30 per diploid genome of M. truncatula. Tnt1 is likely to transpose very actively as we have not detected plants containing only one complete copy of the T-DNA or only one copy of Tnt1 among the 41 T0 lines tested by Southern blot. The multicopy behaviour of the vector used to introduce Tnt1 in M. truncatula may be responsible for the high copy number of the element in this plant. Interestingly, the number of transposed Tnt1 copies in transgenic lines of Arabidopsis fluctuated between 0 and 26 (Courtial et al., 2001). The absence of transposed copies in some regenerated Arabidopsis plants might be caused by differences in the regeneration protocols used for these two plants. For example, 2,4-D, that was shown to induce Tnt1 transposition in Arabidopsis (Pauls et al., 1994), is used in our regeneration protocol at a concentration 10 times higher than that in the A. thaliana regeneration protocol.

Tnt1 is transcribed at different levels in different organs of M. truncatula (flowers, leaves and roots). The transcript level is higher in roots than in old leaves and weak in young leaves and flowers. These data correlate with the expression of Tnt1 observed in its original host tobacco or in heterologous hosts like Arabidopsis and tomato (Mhiri et al., 1997; Moreau-Mhiri et al., 1996; Pouteau et al., 1991). As observed in A. thaliana (Feuerbach et al., 1997), the linear and circular Tnt1 DNA intermediates of replication were also detected in old leaves and roots of M. truncatula.

All the insertions present in the primary transformants were detected in the progeny of the transformed plants at a normal ratio, indicating that the insertions observed in the T0 lines did not represent late somatic transpositions but rather early transposition of Tnt1 during the in vitro culture. Despite the transcription and the replication of the element in the primary transformants of M. truncatula, we did not detect new transposition events in the T1 plants similar to what was observed in Arabidopsis transgenic lines (Lucas et al., 1995). This indicates that the transposition of the element seems to be rare under our growth conditions and that the insertions are stable. The expression and replication of Tnt1 in the T1 plants indicates that it is still transcriptionally active and not silenced in this generation.

As a result of the mechanism of transposition that requires a cytoplasmic replication step before reintegration in the genome, the transposed copies of Tnt1 are supposed to be genetically independent. The segregation of 10 Tnt1 insertions in the progeny of two M. truncatula transgenic plants showed that they can effectively be separated genetically after self-pollination. Such segregation will probably be even more efficient after crossing with wild-type plants. In this case, by starting with a plant containing 16 copies of Tnt1, the first backcross will statistically reduce this number to eight copies, the second to four and the third to two copies. In addition, using Southern blots, plants with the lowest Tnt1 copy number can be chosen for each backcross. Thus, we can predict that plants with single-copy insert can be obtained in three backcrosses. This number is relatively low in comparison to populations mutagenized with ethyl methyl sulfonate (EMS). In this case, the number of mutations per plant is probably higher (at least 100 mutations per genome; McCallum et al., 2000) and as they are not tagged, the number of backcrosses necessary to clean the genome is probably higher than that for the Tnt1 population.

Tnt1 as an insertion mutagen in M. truncatula

The number of Tnt1 copies per genome of M. truncatula (15 on average) will be an advantage for M. truncatula mutagenesis because less transgenic lines will have to be produced to saturate the genome, as compared to a T-DNA-tagging strategy where one transgenic plant is needed for each insertion (Scholte et al., 2002). In addition, Tnt1, like TtoI and Tos17 (Courtial et al., 2001; Hirochika, 2001; Okamoto and Hirochika, 2000), exhibits no obvious insertion specificity, indicating that the different copies are dispersed along the genome. Moreover, the probability to insert in a gene is high for Tnt1 in M. truncatula and A. thaliana (present study and Courtial et al., 2001), as was observed for Tto1 and Tos17 (Okamoto and Hirochika, 2000; Yamazaki et al., 2001). This makes this retrotransposon gene-tagging strategy probably more efficient than T-DNA tagging (Krysan et al., 2002; Scholte et al., 2002). In agreement with the high frequency of gene disruption by Tnt1, we have already observed mutant plants with various phenotypes including albinos, dwarf and flower homeotic mutants in the progeny of the tnk lines described in this work. Genetic analysis of these plants will confirm that these phenotypes result from Tnt1-mediated gene disruption. However, the insertion tnk48-5 described in this work (Figure 4) segregated with a lethal mutation because homozygous plants for this insertion were not detected among the 45 plants analysed. As this insertion occurred in a gene expressed during flower development, we can speculate that in this line Tnt1 inserted in a gene required for proper seed development in M. truncatula, and thus the observed phenotype (lethality) resulted from the disruption of this gene.

Retrotransposons can inactivate genes by disruption, but they can also deregulate them as their LTR contains enhancer sequences, thus allowing activation tagging in the tissues expressing them (Awasaki et al., 1996). In fact, we showed here that the LTR promoter does function efficiently in old leaves and roots of M. truncatula. Thus, Tnt1 may also be used for activation tagging in M. truncatula roots.

We succeeded to re-mobilize Tnt1 by in vitro culture and demonstrated that the infiltration procedure, the Agrobacterium invasion and the antibiotic selection in the medium are not required to induce transposition in M. truncatula, reducing the different in vitro steps susceptible to induce the appearance of somaclonal variants. Testing other conditions for in vivo activation of Tnt1, such as chemical spraying of flowers, may further reduce the tissue culture-induced somaclonal mutations and make the construction of a large Mtruncatula mutant collection easier.

As a result of the very efficient regeneration protocol developed for line R108 in the laboratory, a large number of transgenic plants could be tested for transposition in this line. However, introduction of Tnt1 in the Jemalong J5 line is now feasible and will be important for future genetic studies in order to make allelic tests with mutants already isolated in this line. Using a Tnt1–gus fusion, we have demonstrated here that the transcriptional regulation of Tnt1 is conserved in Jemalong. This result suggests that it will also transpose in the Jemalong line during the early steps of the regeneration process, thus allowing the generation of a Tnt1-tagged mutant library in this line.

Experimental procedures

Plant material and plant growth conditions

Medicago truncatula (Gaertn.) line R-108-1 (c3) has been described by Trinh et al. (1998) and called R108 throughout the text. Plants were grown in soil in the greenhouse and watered alternately by de-ionized water and a nutritive solution (Soluplant, N/P/K: 18/6/26, Duclos International, Lunel Viel, France) at 60% relative humidity under a 16 h photoperiod at 25°C and 15°C for day and night, respectively.

For seed sterilization, seeds were scarified with sand paper, sterilized for 30 min in a Bayrochlor solution (7 g l−1, Bayrochlor; BAYROL GMBH, Germany), rinsed four times in sterile water and germinated in the dark at room temperature on wet Whatman 3 mm paper, in Petri dishes.

Plant transformation and selection

Plant transformation was carried out according to Trinh et al. (1998). In vitro kanamycin selection was carried out by incorporating 40 mg l−1 of this antibiotic in the SHMab medium (Trinh et al., 1998).

Bacterial strain and T-DNA vectors

The Escherichia coli strain DH5a (Sambrook et al., 1989) was used for cloning and the propagation of different vectors. A. tumefaciens EHA105 strain (Hood et al., 1993) was used in all plant transformation experiments. Plasmids were introduced in this strain by triparental mating as previously described (Ratet et al., 1988).

The Tnk23 vector is a derivative of the pBIN19 vector carrying the autonomous Tnt1 retroelement (Lucas et al., 1995). The pHLV5501 vector is also derived from pBin19 and carries a translational LTR-gus-intron fusion (Courtial et al., 2001).

The pMSI vector (Scholte, unpublished results) is a derivative of the pGKB5 vector (Bouchez et al., 1993) carrying a promoterless gus gene, which contains the portable ST-LS1 intron (Vancanneyt et al., 1990).

The p35S-gusI vector is a derivative of the pGPTV-BAR construct (Becker et al., 1992) carrying the 35Sgus-intron reporter gene also derived from the construct described by Vancanneyt et al. (1990).

β-glucuronidase (GUS) assays

Histochemical staining for GUS activity on all plant fragments was carried out according to Scholte et al. (2002).

Isolation of genomic DNA and Southern blot analysis

Plant genomic DNA was extracted from leaves as described by Dellaporta et al. (1983). Between 5 and 10 µg of each DNA was digested with restriction enzymes under conditions specified by the suppliers, separated on 0.6% TAE 1× agarose gels overnight at IV/cm. Southern blot analysis was performed as described by Church and Gilbert (1984). The nptII probe consisted of a PCR fragment of 699 bp located between nucleotides 51 and 750 downstream of the nptII translation initiation site (Beck et al., 1982). The tntA-B probe consisted of a PCR fragment of 1211 bp located between nucleotides 1216 and 2427 of Tnt1 (GenBank #: X13777). The tntC-D probe consisted of a PCR fragment of 879 bp located between nucleotides 3366 and 4245 of Tnt1. The tntG-H probe consisted of a PCR fragment of 609 bp located between nucleotides 626 and 1235 of Tnt1.

Isolation of total RNA and Northern blot

Total RNA from M. truncatula was extracted from 1 g of plant material as described by Kay et al. (1987), and 5 µg was used for Northern blots. Total RNA from leaves of Tnt1 containing A. thaliana plants induced by CuCl2 (Mhiri et al., 1997) was provided by Béatrice Courtial. Northern blots were performed according to Sambrook et al. (1989), and the hybridization performed was similar to the Southern blot.

Polymerase chain reaction

The oligonucleotides ltr1 5′-ATGTCCATCTCATTGAAGAAGTA-3′ and ltr2 5′-GGGAATAAACCCCTTACCAAAA-3′ were used to amplify the LTR of the retrotransposon Tnt1. The oligonucleotides Km5 5′-GAGGCTATTCGGCTATGACTG-3′ and Km3 5′-ATCGGGAGCGGCGATACCGTA-3′ were used to amplify a region of the nptII gene of the tnk23 T-DNA construct. PCR reactions (94°C, 30 sec; 60°C, 30 sec; 72°C, 1 min, 30 cycles) were performed according to the supplier with 0.0125 unit of Taq polymerase µl−1 (Eurobio, France).

Characterization of Tnt1 borders

Tnt1 borders were isolated using two methods: the PCR-walk technique as described by Scholte et al. (2002) and the I-PCR technique (Lucas et al., 1995).

For the PCR-walk technique, the oligonucleotide adaptor consisted of the oligonucleotide 5′-CTAATACGACTCACTATAGGGCTCGAGCGGCCGCCCGGGGAGGT-3′ annealed to the octamer oligonucleotide 5′-ACCTCCCC-3′, with 5′ end phosphorylated and 3′ end aminated. Oligonucleotides 5′-GGATCCTAATACGACTCACTATAGGGC-3′ (Ap1) and 5′-CTATAGGGCTCGAGCGGC-3′ (Ap2) were used in the first and second PCR reactions, respectively, in combination with Tnt1-specific oligonucleotides. For Tnt1 (GenBank #: X13777), the first and second oligonucleotides used for amplification of the 3′ region were LTR3 5′-AGTTGCTCCTCTCGGGGTCG-3′ and LTR4 5′-TACCGTATCTCGGTGCTACA-3′, respectively. Briefly, the technique was as follows: 1 µg genomic DNA was digested overnight by restriction enzyme EcoRV in a final volume of 100 µl. The digestion was extracted using phenol/chloroform and ligated overnight at 16°C in 20 µl final volume in the presence of the adaptor at a final concentration of 5 µm and 10 units of T4 ligase. The ligation was diluted 25 times in water and 2 µl was used for the first PCR reaction (94°C, 45 sec; 72°C, 2 min, 30 cycles). This first PCR reaction was diluted 100 times in water, and 1 µl was used for the second PCR reaction (for details, see Scholte et al., 2002).

For the I-PCR technique, 500 ng genomic DNA was digested with EcoRI and MfeI overnight at 37°C and precipitated after heat inactivation of the enzyme. Ligation of the fragments was performed overnight at room temperature in 200 µl ligation buffer (Promega, France) with 8 units of ligase (Promega). After heat inactivation of the enzyme and precipitation, the final products were re-suspended in 20 µl of water. Five microlitres were used for the first PCR reaction (94°C, 30 sec; 72°C, 2 min, 30 cycles) with the oligonucleotides LTR3 5′-AGTTGCTCCTCTCGGGGTCG-3′ and LTR5 5′-GCCAAAGCTTCACCCTCTAAAGCCT-3′. Then, 1 µl of the first reaction product, diluted 100 times, was used for the second PCR reaction (94°C, 30 sec; 60°C, 30 sec; 72°C, 2 min, 30 cycles) with the oligonucleotides LTR4 5′-TACCGTATCTCGGTGCTACA-3′ and LTR6 5′-GCTACCAACCAAACCAAGTCAA-3′. PCR reactions were performed according to the supplier's instructions in a volume of 20 µl for the first reaction and 50 µl for the second, with 0.0125 unit of Taq polymerase µl−1 (Eurobio, France).

The PCR fragments amplified by PCR-walk or I-PCR were cloned into the pGEM-Teasy vector (Promega) and sequenced.

Sequencing reaction and sequence analysis

DNA sequences were determined on double-stranded DNA using the BigDye terminator v3.0 kit (Applied Biosystem, USA) following the instructions of the supplier. The sequences of the tagged loci were compared to the sequences of the database using the blast program http://blast.genome.ad.jp/ or the tigrMedicago gene index http://www.tigr.org/tdb/mtgi.

Acknowledgements

We are very grateful to Marije Scholte for providing the pMS1 vector, Béatrice Courtial for providing the RNA sample of Tnt1 transgenic A. thaliana treated with CuCl2, Carine Sicot for help in tissue culture and Nathalie Mansion for photographic work. Isabelle d'Erfurth was supported by a French MENRT doctoral fellowship and the European project Medicago (QLG2-CT-2000–00676).

EMBL sequence accession numbers: AJ511822, AJ511823, AJ511824, AJ511825, AJ511826, AJ511827, AJ511828, AJ511829, AJ511830, AJ511831, AJ511832, AJ511833, AJ511834, AJ511835, AJ511836, AJ511837, AJ511838, AJ511839, AJ511840, AJ511841, AJ511842, AJ511843, AJ511844, AJ511845, AJ511846, AJ511847, AJ511848, AJ511849, AJ511850, AJ511851, AJ511852, AJ511853, AJ511854, AJ511855, AJ511856, AJ511857, AJ511858, AJ511859.

Ancillary