Temperature controls nuclear import of Tam3 transposase in Antirrhinum


  • Kaien Fujino,

    1. Laboratory of Crop Physiology, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
    Search for more papers by this author
    • Present address: Biotechnology Sector, Central Research Institute of Electronic Power Industry (CRIEPI), 1646 Abiko, Chiba 270-1194, Japan.

  • Shin-nosuke Hashida,

    1. Laboratory of Genetic Engineering, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
    Search for more papers by this author
    • Present address: Biotechnology Sector, Central Research Institute of Electronic Power Industry (CRIEPI), 1646 Abiko, Chiba 270-1194, Japan.

  • Takashi Ogawa,

    1. Laboratory of Plant Breeding, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
    Search for more papers by this author
  • Tomoko Natsume,

    1. Laboratory of Plant Breeding, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
    Search for more papers by this author
  • Takako Uchiyama,

    1. Laboratory of Plant Breeding, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
    Search for more papers by this author
  • Tetsuo Mikami,

    1. Laboratory of Genetic Engineering, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
    Search for more papers by this author
  • Yuji Kishima

    Corresponding author
    1. Laboratory of Plant Breeding, Research Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan
      (fax +81 11 706 4934; e-mail kishima@abs.agr.hokudai.ac.jp).
    Search for more papers by this author

  • These authors contributed equally to this work.

(fax +81 11 706 4934; e-mail kishima@abs.agr.hokudai.ac.jp).


It has been proposed that environmental stimuli can activate transposable elements (TEs), whereas few substantial mechanisms have been shown so far. The class-II element Tam3 from Antirrhinum majus exhibits a unique property of low-temperature-dependent transposition (LTDT). LTDT has proved invaluable in developing the gene isolation technologies that have underpinned much of modern plant developmental biology. Here, we reveal that LTDT involves differential subcellular localization of the Tam3 transposase (TPase) in cells grown at low (15°C) and high (25°C) temperatures. The mechanism is associated with the nuclear import of Tam3 TPase in Antirrhinum cells. At high temperature, the nuclear import of Tam3 TPase is severely restricted in Antirrhinum cells, whereas at low temperature, the nuclear localization of Tam3 TPase is observed in about 20% of the cells. However, in tobacco BY-2 and Allium cepa (onion) cells, Tam3 TPase is transported into most nuclei. In addition to three nuclear localization signals (NLSs), the Tam3 TPase is equipped with a nuclear localization inhibitory domain (NLID), which functions to abolish nuclear import of the TPase at high temperature in Antirrhinum. NLID in Tam3 TPase is considered to interact with Antirrhinum-specific factor(s). The host-specific regulation of the nuclear localization of transposase represents a new repertoire controlling class-II TEs.


Under unfavorable environmental conditions, plants immediately respond in a physiological way to avoid fatal consequences, but plants also alter the genome per se to produce potential progenies adapted to the given environmental changes. Mechanisms by which the physiological responses in plants follow the environmental changes have been intensively studied in various aspects. It is known that the structural changes of the plant genomes are occasionally induced by transposable elements (TEs) activated by environmental stresses (Grandbastien, 1998; Kalendar et al., 2000), whereas the mechanisms regarding the activation processes of TEs are limited in our knowledge. To maintain genome integrity, plants may adopt diverse strategies to regulate the activities of transposable elements.

Tam3 from Antirrhinum majus is a class-II transposon belonging to the hAT superfamily (Calvi et al., 1991). Tam3 excision and, by implication, transposition, exhibit the unusual features of activation at low growth temperatures (15°C) and inhibition at high temperatures above 25°C (Harrison and Fincham, 1964; Carpenter et al., 1987) (Figure 1a). Activation of Tam3 can occur as a rapid and reversible response to a temperature change during the lifetime of a single plant (Hashida et al., 2003). Such low-temperature-dependent transposition (LTDT) of Tam3 represents a response of transposon to environmental stimuli. Because the activity of Tam3 can be controlled by temperature, this transposon system has been used for the generation of mutations and a gene tagging system that have made remarkable contributions to the study of developmental and metabolic systems in plants (Coen et al., 1989; Schwarz-Sommer et al., 2003).

Figure 1.

 Phenotypic changes in Antirrhinum flowers as a result of variegation in plants grown at 25° and 15°C, and detection of Tam3 TPase.
(a) The different variegated flower phenotypes at 25° and 15°C resulted from the low-temperature-dependent transposition (LTDT) of Tam3 in the pallidarecurrens::Tam3 allele of Antirrhinum.
(b) Immunoblot analysis to detect the Tam3 TPase using an antipeptide antibody, pN1, against amino acid positions 212–230 of Tam3 TPase (see Figure S1). Total proteins were extracted from the leaves of Antirrhinum plants grown at 25° or 15°C. The antibody pN1 specifically recognizes a peptide of 100 kDa, which corresponds to the 803-aa peptide of Tam3 TPase (Uchiyama et al., 2008).

Unlike known mechanisms in the stress responses of other TEs, LTDT is not regulated at the level of transcription of the TPase gene (Hashida et al., 2003). Although a reversible change of DNA methylation state in Tam3 that occurred in somatic tissues was positively correlated with growth temperature (Hashida et al., 2003), methylation was not a cause of LTDT of Tam3 (Hashida et al., 2005). Instead, nuclear extracts from plants grown at 15°C were capable of binding to the Tam3 sequence, but the extract from plants grown at 25°C did not bind, suggesting that the methylation change in Tam3 is affected by the binding ability of its transposase (TPase) to Tam3, which is dependent on the growth temperature (Hashida et al., 2006). Previous results in vivo led to a conclusion that Tam3 TPase exhibits differential binding abilities to Tam3 between 15° and 25°C that might be related to LTDT.

Here, it is reported that the nuclear localization of Tam3 TPase in Antirrhinum is arrested at temperatures at which Tam3 transposition is inhibited. The Tam3 TPase reporter constructs fused with green fluorescence protein (GFP) and the luciferase (LUC) assay demonstrated that the nuclear localization of Tam3 TPase in Antirrhinum cells is severely restricted at high temperature, and is permitted in about 20% of the cells at low temperature. The results are provided to demonstrate that LTDT of Tam3 is controlled by the nuclear localization of TPase, and evidence is shown for specific interactions between Tam3 TPase and host factor(s) to establish LTDT.


Presence of Tam3 TPase in Antirrhinum cells

The LTDT of Tam3 is not regulated by transcription of the Tam3 TPase gene or the DNA methylation state of Tam3 (Hashida et al., 2003, 2005). To test whether LTDT is correlated with the level of the Tam3 TPase in Antirrhinum cells, immunoblot analysis was performed using an anti-peptide antibody, pN1 (Uchiyama et al., 2008) (see Figure S1). pN1 detected a polypeptide of about 100 kDa corresponding to Tam3 TPase in young leaf proteins extracted from Antirrhinum plants grown at 15° and 25°C (Figure 1b). This result demonstrated that the processes prior to translation of TPase are not responsible for LTDT.

Because TPases should be transported into the nucleus in order to catalyze the transposition reaction (Heinlein et al., 1994; Ono et al., 2002), it was considered that LTDT might reflect different accessibility of the TPase to the Tam3 DNA sequences at the two temperatures: the low temperature could enhance the TPase accessibility to the Tam3 sequences, whereas the high temperature could abolish the accessibility. In previous results, proteins in a nuclear extract from plants grown at 15°C, but not at 25°C, bind to the Tam3 sequence (Hashida et al., 2006). Differential nuclear localization of the TPase at high and low temperatures is thus a possible mechanism to explain LTDT.

Nuclear localization of Tam3 TPase

To confirm the above assumption, transient expression analyses were conducted using a plasmid, pTam3TPase-GFP, containing the Tam3 TPase gene fused to the GFP gene. The plasmid was introduced into young petal epidermal cells of Antirrhinum plants grown at 15° or 25°C using a particle bombardment method, and the bombarded petals were further incubated for 24 h at 15° or 25°C (Figure 2). The continuous low temperature (Ll) or low temperature after bombardment (Hl) conditions did not allow the plasmid to express TPase-GFP in the Antirrhinum petals, because the transient expression of the plasmid required relatively high temperatures. Continuous high temperature (Hh) conditions allowed the cells to express GFP, and most expressed GFP in the cytoplasm, whereas only 1% of the cells expressed GFP in their nuclei. After bombardment, in high-temperature (Lh) conditions, GFP was detected in the nuclei of about 18% of the cells with green fluorescence (Figure 2). These results showed that the TPase-GFP was not directed to the nucleus in most cells of the Antirrhinum plants grown at high temperature, whereas about 20% of the cells localized TPase-GFP to the nuclei at low temperature. When pTam3TPase-GFP was delivered into tobacco BY-2 cells and onion epidermal cells, GFP signals were observed in the nuclei of nearly 100% of the cells (see Figure 5 and Figure S2), suggesting that LTDT of Tam3 is an Antirrhinum-specific phenomenon.

Figure 2.

 Subcellular localization of Tam3 TPase at different temperatures.
Plasmids were delivered into Antirrhinum petal cells by the bombardment method. Before introduction of the plasmid, the plants were grown at 25° (H) or 15°C (L) for at least 3 weeks. The young petals bombarded with the plasmid were incubated at 25° (h) or 15°C (l) for 24 h, and the subcellular localization of GFP was observed by fluorescence microscopy. Petal cells bombarded with a plasmid carrying only the GFP gene showed preferential nuclear localization of the GFP protein, whereas in the case of pTam3TPase-GFP, two GFP localization patterns were observed: one in which GFP was present in the nucleus and the other in which it was absent from the nucleus. The histogram shows the frequencies of cells with nuclear localization among those expressing GFP; N.D., GFP was not detected. Scale bars: 10 μm.

Figure 5.

 Effect of leptomycin B (LMB) on nuclear localization of Tam3 TPase.
Each of the plasmids, pTam3TPase-GFP or pJ4-RSG-GFP (RSG), was introduced into protoplasts prepared from Antirrhinum leaves and tobacco BY-2 cells grown at 25°C, and the protoplasts were incubated with 20 ng ml−1 of LMB (+) or without LMB (−) for 24 h at 25°C (J). For the control, pTam3TPase-GFP was delivered into leaf protoplasts from Antirrhinum leaves grown at 15°C. Representative images of the subcellular localizations of GFP fusion proteins in Antirrhinum leaf protoplasts (a–d) and tobacco BY-2 protoplasts (e–h).
(a) The leaf protoplasts from Antirrhinum plants grown at 15°C show GFP signals derived from pTam3TPase-GFP in the nucleus.
(b) Leaf protoplasts from Antirrhinum plants grown at 25°C show GFP signal derived from pTam3TPase-GFP in the cytoplasm.
(c) Leaf protoplasts from Antirrhinum plants grown at 25°C show GFP signals derived from pJ4-RSG-GFP in the cytoplasm.
(d) Leaf protoplasts from Antirrhinum plants grown at 25°C, which were treated with LMB, show GFP signals derived from pJ4-RSG-GFP in the nucleus.
(e) Protoplasts from BY-2 cells cultured at 25°C show GFP signals derived from pTam3TPase-GFP in the nucleus.
(f) Protoplasts from BY-2 cells cultured at 25°C, treated with LMB, show GFP signals derived from pTam3TPase-GFP in the nucleus.
(g) Protoplasts from BY-2 cells cultured at 25°C show GFP signals derived from pJ4-RSG-GFP in the cytoplasm.
(h) Protoplasts from BY-2 cells cultured at 25°C, treated with LMB, show GFP signals derived from pJ4-RSG-GFP in the nucleus.
(i) Microscopic field image indicating multiple Antirrhinum protoplasts that show GFP signals derived from pJ4-RSG-GFP. Scale bars: 10 μm.
(j) The histogram indicates the percentage of cells expressing GFP in the nucleus among cells with GFP signals. Standard deviations were obtained from three repetitions.

Previously, Antirrhinum callus showed LTDT of Tam3 similar to the plants (Hashida et al., 2005). Then, we examined transgenic calli of Antirrhinum leaf that produced a chimera protein of Tam3 transposase and GFP. The plasmid pGWB5-Tam3 that contains the Tam3 TPase gene in the pGWB5 construct consisting of 35S promoter and GFP gene (Figure 3a) was introduced into the Antirrhinum genome by Agrobacterium-mediated transformation. In contrast to the transient expression of GFP, the transgenic calli stably expressed the fluorescence even at a continuous low temperature of 15°C (Figure 3b). The nucleus-to-cytoplasm GFP (NCG) ratio was calculated based on the GFP fluorescent intensities of the nucleus and cytoplasm in the cell. The Antirrhinum calli transformed with the plasmid pGWB5 encoding merely GFP did not exhibit any difference in distributions of the NCG ratio between cells grown under low and high temperatures (Figure 3c). In the calli expressing GFP-fused TPase, however, the NCG ratios of the cells grown under low temperatures were clearly distributed in a higher ratio compared with that of cells grown under high temperatures, which was similar to the findings observed in calli transformed with pGWB5 (Figure 3c). Therefore, the transgenic Antirrhinum calli verified that the low temperature induces the GFP-fused TPase to accumulate in the nuclei.

Figure 3.

 Differential localization of the Tam3 TPase in transgenic Antirrhinum calli at different temperatures.
(a) The construction of the plasmid pGWB5-Tam3 was accomplished by replacement of the ccdB gene in pGWB5 with the Tam3 TPase open reading frame (ORF) using GATEWAY Technology (Invitrogen). P35S; CaMV 35S promoter, B1 and B2; attB1 and attB2 as the sites for BP recombinase, R1 and R2; attR1 and attR2 as the sites for LR recombinase, TNO; TNOS, the terminater sequence of Nopalin syntase gene.
(b) The two constructs (pGWB5-Tam3 and pGWB5) were introduced into the Antirrhinum calli, which were grown at 15° or 25°C. The GFP fluorescence was detected in the cells grown at both temperatures, but strong localization at the nuclei was observed in the calli carrying pGWB5-Tam3 grown at 15°C. Nuclear localization was confirmed by 4′,6-diamidino-2-phenylindole (DAPI) staining. Scale bars: 10 μm.
(c) The nucleus-to-cytoplasm GFP (NCG) ratios for GFP fluorescence intensities were estimated in 50 cells randomly selected in each plot, and the frequencies (number of cells) were depicted in two histograms for the calli containing pGWB5-Tam3 and pGWB5. The blue bar indicates cells grown at 15°C, whereas the red bar indicates cells grown at 25°C.

Quantitative measurements of nuclear localization of Tam3 TPase

For quantitative measurement of the level of Tam3 TPase in nuclei, the experiments using the GAL4 system were adopted (Aoyama et al., 1995; Igarashi et al., 2001). This system uses the effector protein containing GAL4 DNA-binding domain (GAL4 BD) and VP16 activator domain (VP16), both of which are required for the transcriptional activation of the reporter gene carrying GAL4 binding sites in its promoter sequence. We constructed two effector plasmids, pGATam3 and pGAGUS, where the Tam3 TPase or GUS proteins were fused to GAL4 BD and VP16 activator domains (Figure 4a). The reporter plasmid, pGBL, contained two operons for Renilla luciferase and firefly luciferase genes, as shown in Figure 4a: the Renilla luciferase gene driven by the 35S promoter is a reference to standardize differences in efficiency of the bombardment delivery, and the transcription of the firefly luciferase gene can be induced by the binding of the effector protein to the GAL4 binding sites (six tandem copies). The firefly luciferase is produced by nuclear import of the effector proteins containing GAL4 BD and VP16. The constructs were introduced into young petal epidermal cells of Antirrhinum plants grown at 25°C by the bombardment method, and the bombarded petals were incubated for 16 h at 15° or 25°C, and then the firefly and Renilla luciferase activities were measured. The outcome for the level of nuclear import of Tam3 TPase or GUS protein was indicated by the ratio of the firefly luciferase activity to the Renilla luciferase activity (LUC ratio). When pGATam3 was bombarded along with pGBL, the incubation at 15°C of the petal cells showed as much as five times higher LUC ratio relative to the incubation at 25°C (Figure 4b). The delivery of pGAGUS and pGBL into the petal cells resulted in the higher LUC ratio at 15°C relative to that at 25°C, but it was not statistically significant (Figure 4b). In the case where pGBL alone was introduced, the LUC ratios were much reduced in both of the temperature conditions (Figure 4b). It is noteworthy that GAL4 BD in the effector protein contains an NLS, and so is the effector protein directed to the nucleus. Accordingly, the GUS fusion protein from the pGAGUS construct should be fully imported into nuclei, as pGAGUS induced high LUC ratios. The LUC ratio in the pGATam3-introduced cells incubated at 15°C were about one-fifth of that of pGAGUS-introduced cells incubated at 15°C (Figure 4b). This was the same proportion with that estimated with pTam3TPase-GFP: about 20% of the cells with Tam3 TPase-GFP had the signal in nuclei at low temperature (Figure 2). Therefore, although low temperature facilitated nuclear localization of Tam3 TPase in Antirrhinum, about 80% of Tam3 TPase might not be permitted to be present in Antirrhinum nuclei. At high temperature, nuclear localization of Tam3 TPase was even more strictly restricted. The quantitative measurements using the luciferase gene constructs verified the results in the analyses using the GFP fusion proteins.

Figure 4.

 Quantitative analysis of nuclear localization of Tam3 TPase using the GAL4 system.
(a) Two effector and one reporter plasmids were constructed as described in the Experimental procedures. Both the effector plasmids, pGATam3 and pGAGUS, commonly contain CaMV 35S and omega motif (Ω) for the promoter sequences, and the terminater sequence of the Nopalin syntase gene (TNOS). The coding sequences in pGATam3 and pGAGUS also shared the GAL4 DNA binding domain (GAL4 BD) and VP16 activator domain (VP16), which were fused to Tam3 TPase or GUS polypeptides, respectively. The reporter construct, pGBL, consists of the firefly LUC and Renilla LUC genes. The firefly LUC gene is induced by the binding of GAL4 BD plus the VP16 domain to six tandem motifs of the GAL4 binding site, whereas the Renilla LUC gene is constitutively transcribed with the CaMV 35S promoter sequence.
(b) Measurements of the luciferase activities were carried out with Antirrhinum petal cells, which were incubated for 16 h at 25° or 15°C after bombardment of the plasmids. LUC activity is the ratio of firefly LUC activity to Renilla LUC activity (×1000), and indicates a level of nuclear localization of Tam3 TPase. *LUC ratios between 25° and 15°C were significantly different at the 1% level (t = 0.0003); **LUC ratios between 25° and 15°C were not significantly different at the 5% level (t = 0.116).

Arrest of nuclear import of Tam3 TPase at high temperature

As described in the previous sections, two mechanisms for the temperature-dependent differential nuclear localization of TPase were considered: (i) TPase is imported into the nucleus only at low temperatures; (ii) TPase is exported from the nucleus into the cytoplasm at high temperatures. To test these possibilities, leptomycin B (LMB), which inhibits the export of nuclear proteins into the cytoplasm by targeting exportin 1, was applied to protoplasts prepared continuously at 25°C. The effect of LMB was verified using the plasmid pJ4-RSG-GFP, which encodes a GFP-fusion protein containing an NLS along with a nuclear export signal (NES) (Igarashi et al., 2001). The RSG-GFP fusion protein was predominantly localized to the cytoplasm of both Antirrhinum and BY-2 protoplasts (Figure 5c,g,j). In protoplasts treated with 20 ng ml−1 LMB, the ratios of cells that accumulated RSG-GFP protein in the nucleus to the number of total cells expressing GFP increased in both Antirrhinum (from 28 to 88%) and BY-2 (from 15 to 53%) protoplasts (Figure 5d,h,j). In the same protoplasts, LMB failed to alter the ratios for the nuclear localization of GFP from pTam3TPase-EGFP, which was mostly observed in the cytoplasm of Antirrhinum and in the nuclei of BY-2 cells (Figure 5f,j). Therefore, the temperature-dependent differential nuclear localization of TPase must result from the inhibition of TPase import into the nucleus in cells grown at 25°C, and not from the inhibition of export. The control of nuclear import of TPase underlies the LTDT of Tam3 in Antirrhinum. Even under low temperature conditions, the nuclear translocation of Tam3 TPase-fusion protein was limited to about 20% of the total cells expressing GFP or LUC in Antirrhinum (Figures 2 and 3). This may be a mechanism to suppress transposon activity and maintain genome stability that has evolved in Antirrhinum.

Nuclear localization signals in Tam3 TPase

Nuclear localization signals are intrinsic for the entry of nuclear proteins to the nucleus after translation in the cytoplasm. Based on the amino acid sequence of the Tam3 TPase, three short stretches (approximately 35 amino acids) were predicted as NLSs, which contain six or seven basic amino acid residues (K/R) (Figure 6 and Figure S1). Each of the three regions was fused to a GFP plus GUS protein (to prevent GFP from autonomously entering the nucleus). Bombardment with every construct resulted in green fluorescence that was mostly detected in the nuclei in both Antirrhinum petals and onion epidermal cells incubated at 25°C (Figure 6). The results revealed that Tam3 TPase possesses at least three NLSs, which facilitated entry into the nucleus. These findings were supported by other data obtained using onion cells bombarded with deletion mutants of the Tam3 TPase (Figure S2). These results suggest that the Tam3 TPase is equipped with a domain for the arrest of nuclear translocation of the TPase at high temperatures.

Figure 6.

 Identification of three nuclear localization signals (NLSs) in Tam3 TPase.
Three acidic domains of Tam3 TPase were tested as NLS candidates. Each of the candidate domains was fused to the GFP-GUS gene construct, and the resultant plasmids were introduced into onion epidermal cells or Antirrhinum petal cells at 25°C (Hh). The GFP-GUS fusion protein could not be transported to the nucleus, but each NLS could direct the GFP-GUS protein to the nucleus in most of the cells. The positions and amino acid sequences of the three NLSs are indicated below the panels and are listed in Figure S1. Scale bars: 30 or 10 μm for onion and Antirrhinum panels, respectively.

Nuclear localization inhibitory domain in Tam3 TPase

To test whether Tam3 TPase contains a domain required for the arrest of nuclear translocation at high temperatures, further transient assays were carried out. One intact and four truncated sequences of the Tam3 TPase gene fused to the GFP gene were introduced into the young Antirrhinum petals of plants grown at 15° or 25°C through the bombardment method (Figure 7, left). The bombarded tissues were incubated in Hh or Lh conditions as described in Figure 2. The truncated constructs contained at least one NLS, so that the GFP fusion proteins potentially entered the nuclei, as depicted in Figure 7. When the two plasmids pTP231–577E and pTP578–803E were introduced into the petals of plants grown at 15° and 25°C, most of all the cells (87.5–100% of cells with GFP signal) expressed green fluorescence in the nuclei under either of the temperature conditions (Figure 7, right). The GFP fusion protein from pTam3TPase-GFP was delivered into the nuclei in 1 and 18% of the cells observed in Hh and Lh conditions, respectively (Figures 2 and 5). Similarly, low frequencies of nuclear localization of GFP signal were observed in the assays with pTP1–577E and pTPΔBE (Figure 7). These three constructs, pTam3TPase-GFP, pTP1–577E and pTPΔBE, all encoded the amino acid sequences 55–244 of the Tam3 TPase (Figure 7). This 55–244 amino acid sequence should be associated with the arrest of nuclear translocation of Tam3 TPase itself. Therefore, this region was designated as the nuclear localization inhibitory domain (NLID), which is potentially capable of suppressing the translocation of Tam3 TPase to the nuclei in Antirrhinum cells. Compared with pTP231–577E and pTP578–803E, these three constructs had even lower frequencies of nuclear localization of the GFP signal in both the low and high temperature conditions. The NLID of Tam3 may not only function at high temperature to strictly prevent nuclear translocation of the TPase, but also functions at low temperature to confine nuclear translocation to around 20% of the total TPase yield.

Figure 7.

 Identification of the nuclear localization inhibitory domain (NLID) of Tam3 transposase.
The left panel shows the intact and truncated constructs derived from the Tam3 TPase gene. Each construct fused with the GFP gene was introduced into young petals of Antirrhinum by the bombardment method. Numbers in the constructs show the amino acid positions in Tam3 TPase. Black stripes indicate the positions of NLSs. pTPΔBE has two deletions of the Tam3 TPase at amino acid positions 1–54 and 245–588. The right panel presents a histogram with the proportions of cells with nuclear localization of GFP compared with the total number of cells with green fluorescence, and also by fractions in the columns (numerator, number of the cells showing nuclear GFP localization; denominator, total number of cells observed). Each of the constructs was introduced to young petals of plants grown at 15° (L) or 25°C (H) for at least 3 weeks, and the bombarded petals were incubated at 25°C (h) for 24 h; Lh, white bars; Hh, gray bars.

These five plasmids used in the above analysis were delivered into onion epidermal cells by the bombardment method, and most of the cells with GFP clearly exhibited green fluorescence in nuclei (Figure S2). The NLID may interact directly with Antirrhinum-specific factor(s) to prevent nuclear translocation of the TPase itself. The NLID is considered as a Tam3-unique region; no homology was detected in the N-terminal 170 amino acids of Tam3 TPase, which partly comprises the NLID, although the rest of the TPase amino acid sequence was homologous with rice putative TPases (Dart) (AB206820) (Fujino et al., 2005; Tsugane et al., 2006).


This study elucidated the temperature-dependent differential nuclear localization of Tam3 TPase: low temperatures allowed the TPase to enter the nuclei, but high temperatures did not allow entry. Control of the nuclear translocation of the TPase thus underlies the LTDT of Tam3 in Antirrhinum. The regulatory systems of transposon activities identified so far are known to have specific limitations: the germ line-specific transposition of the P element in Drosophila because of splicing being limited to the germ cells (Laski et al., 1986); the inactivation of Tc1 elements in the germ line of Caenorhabditis elegans attributable to RNA interference (Ketting et al., 1999); and the developmentally regulated transposition of Zea mays (maize) transposons (Levy and Walbot, 1990). Repressor genes have also been identified against P (Black et al., 1987), Spm (Cuypers et al., 1988) and Mu (Slotkin et al., 2005) transpositions, which result from aberrant transposase genes or the rearrangements of their derivative elements. In addition, non-specific regulation of transposon activity has been observed because of DNA methylation and/or RNA silencing (Martienssen, 1998; Colot and Rossignol, 1999; McDonald et al., 2005; Buchon and Vaury, 2006). LTDT of Tam3 occurring at a post-translational stage is considered as a unique regulatory mechanism relative to the above regulation mechanisms of transposon activity.

For Tam3 TPase, the differential subcellular localization regulated by temperature is unlikely to be accomplished by itself, because the same phenomenon did not arise when pTam3TPase-GFP was introduced into tobacco BY-2 and onion cells, where green fluorescence was preferentially detected in the nuclei at high temperature. The control of nuclear localization of Tam3 TPase should be established by cooperation with Antirrhinum host factor(s). Martin et al. (1989) and Haring et al. (1991) also showed no enhancement of Tam3 transposition at low temperatures in transgenic tobacco plants. These previous results support the prediction that LTDT requires Antirrhinum-specific factor(s).

In general, TPases are transported from the cytoplasm into the nucleus after translation, because of the presence of NLS-related domains, to allow TPases to bind to nuclear DNA for the cut-and-paste processes. Thus, NLSs seem to be essential for TPases. The two transposons, Ac (Boehm et al., 1995) and Mu (Ono et al., 2002), in maize have at least one major transcript in their autonomous elements encoding proteins containing NLS signals, which are capable of directing the proteins to the nuclei in petunia protoplasts and onion epidermal cells, respectively. Tag1 in Arabidopsis was also inferred to possess an NLS from the amino acid sequence (Liu and Crawford, 1998). At least three monopartite NLSs were identified in Tam3 TPase. When GFP fusion proteins carrying each of the NLSs were introduced into Antirrhinum or onion cells, the green fluorescence was localized to the nuclei at both temperatures. In addition to these monopartite NLSs, Tam3 TPase harbors an NLID that was identified as a region to inhibit nuclear localization of the TPase. The NLID consisted of an amino acid sequence unique to Tam3 TPase, and is a possible region to regulate LTDT of Tam3 and to interact with Antirrhinum-specific factor(s). The NLID might also play a role to limit nuclear translocation up to approximately 20–30% of the total Tam3 TPase production, even at low temperatures. These results seem to reflect the complexity of the host mechanism involved in LTDT.

Tam3 is classified into the hAT superfamily (Calvi et al., 1991), of which related elements have been widely found in multicellular organisms, implying that the hAT superfamily probably arose before the separation of plants, animals and fungi. These elements are considered to have co-evolved vertically with individual host organisms, because there is no evidence for their horizontal transfer (Rubin et al., 2001). The distinct behavior of Tam3 in Antirrhinum might reflect coordinate evolution between host-specific factor(s) and the transposon.

Experimental procedures

Plant materials

We used the Antirrhinum majus line HAM5 (nivearecurrens: Tam3), which gives rise to variegated petals at low temperatures of around 15°C because of Tam3 transposition. This line was initially grown for 2 months at 25°C and subsequently transferred to a 15°C growth chamber or grown continuously at 25°C. The tobacco BY-2 cell line was maintained in the liquid MS medium and subcultured every week. Onions were used within 1 week of purchase.

Antibody production and purification

Antibody production was performed by the Molecular Biology Laboratory (Nagoya, Japan). An oligopeptide, GCGTGTLTRHLTAKHKNRD, which corresponds to the amino acid residues 212–230 of Tam3 TPase (see Figure S1) was synthesized and used to immunize a rabbit. To purify the antibody, the serum obtained was incubated with the resin-immobilized oligopeptide (FMP-activated Cellulofine; Seikagaku Kogyo, http://www.seikagaku.co.jp/)) overnight at room temperature (23°C). The resin was washed extensively with wash buffer containing 20 mm Tris–HCl (pH 7.5), 1 m NaCl and 1% Triton X-100, and then with Tris-buffered saline (TBS). Absorbed antibody molecules were eluted with 0.1 m glycine-HCl (pH 2.5) and neutralized immediately with 1/20 volume of 1 m Tris at 4°C. The purified antibody was designated as pN1.

Immunoblot analysis

Total protein was extracted from the young leaves or flower petals of Antirrhinum plants. About 0.5 g of tissue was homogenized with 1 ml of protein sample buffer containing 60 mm Tris–HCl (pH 6.8), 2% SDS, 10% glycerol, 5 μg of aprotinin and 35 μg of PMSF, and then supplemented with 50 μl of 2-mercaptoethanol. The homogenate was centrifuged for 5 min at 20 400 g. The resulting supernatant was separated from the cell debris. The protein concentrations of the extracts were measured using a Protein Assay kit (Bio-Rad, http://www.bio-rad.com). About 20 μg of protein was separated by 10% (w/v) SDS-PAGE and transferred to a poly-vinylidenedifluoride membrane. After blocking with 4% (w/v) skimmed milk, the membrane was incubated overnight at 4°C with 2500-fold-diluted primary antibody. The membrane was incubated for 1 h with 5000-fold-diluted secondary antibody (anti-rabbit immunoglobulin antibody conjugated to horseradish peroxidase), and then stained with ECL Western Blotting Detection Reagent (RPN2106; Amersham Bioscience, now part of GE Healthcare, http://www.gelifesciences.com) according to the manufacturer’s recommendations. Signals were detected using Fuji X-ray films.

Plasmid construction

To construct plasmids able to express GFP, modifications were made to the basic construct pJB1414, which contains the CaMV 35S promoter, EGFP coding sequence and the terminal sequence of nopaline synthase (TNOS) in a cistron, and the ampicillin resistance gene. For the TPase gene-fusion plasmid, pTam3TPase-GFP, the full-length Tam3 TPase coding sequence (2409 bp in AB038406) was inserted between the 35S promoter and EGFP sequences in pJB1414 (Uchiyama et al., 2008). This insertion process was mediated by the Gateway reaction (Invitrogen, http://www.invitrogen.com) and the primers used in GATEWAY cloning are listed in Table S1. Similarly, the plasmid, pGWB5-Tam3, for Agrobacterium-mediated transformation, was constructed, based on pGWB5 (Nakagawa et al., 2007), by the GATEWAY cloning system. The plasmid, pGWB5-Tam3, locates the Tam3 TPase gene upstream of the synthetic green fluorescent protein with the S65T mutation (sGFP gene), which is present between the 35S promoter and TNOS in pGWB5 (Figure 3a). Deletions of parts of the Tam3 TPase coding sequence for pTP1-577E, pTP231-577E, pTP578-803E and pTP1-351E were produced by PCR, and the deletion of pTPΔBE was performed by self-ligation after BspT1071 digestion with the PCR segment of the Tam3 TPase gene (the sequence corresponding to Tam3 TPase amino acid positions 55–803), which can take away the DNA sequence between amino acid positions 245 and 588 of Tam3 TPase. These truncated sequences were also introduced between the 35S promoter and EGFP coding sequence in pJB1414 through the GATEWAY method, as described above. To survey the NLS sequences, each of the candidate NLSs (Table S2) with a basic amino acid cluster was inserted between the 35S promoter and EGFP sequences in pJB1414. These plasmids also contained the GUS gene, which was inserted between the EGFP coding and TNOS sequences of pJB1414, to prevent autonomous translocation to the nucleus (the EGFP molecule by itself tends to enter the nucleus).

Transient expression assay

Transient expression of GFP in Antirrhinum petal and onion epidermal cells was carried out using the Helium Biolistic gene transformation system (Bio-Rad). The epidermis of each tissue was peeled and placed on half-strength MS medium. Plasmid DNA (approximately 2 μg) was precipitated onto gold particles (1.6 μm in diameter) as described previously (Hashida et al., 2003). The bombardment delivery into intact tissue was performed according to the method of Uchiyama et al. (2008). The subcellular localization of GFP fusion proteins was observed using a UV fluorescence microscope (Olympus, http://www.olympus-global.com) equipped with a UV-A fluorescence filter set (330- to 380-nm excitation filter; 400-nm dichroic mirror; 420-nm barrier filter). Bombardments were repeated three or more times for each plot.

Agrobacterium-mediated transformation and GFP microscopic observation

Leaves from in vitro-grown plants were used for callus induction in MS medium containing 5 mg L−1 Picloram (Sigma-Aldrich, http://www.sigmaaldrich.com). Two weeks after callus induction, 5-mm square pieces of leaf laminae were inoculated with freshly grown Agrobacterium tumefaciens strain LBA4404, transformed with pGWB5-Tam3 or pGWB5 (Figure 3a). Inoculated leaf segments were co-cultivated on 0.2% Gellan gum (Wako, http://www.wako-chem.co.jp) solidified MS medium containing 30 g L−1 sucrose. A total of 150 leaf explants were treated with Agrobacterium and cultured in plastic dishes. After 5 days of co-cultivation, the leaf segments were transferred to the same medium with the addition of 300 mg L−1 cefotaxime (Hoechst AG, now part of Sanofi-Aventis, http://www.sanofi-aventis.de) to eliminate Agrobacterium and the additions of 25 mg L−1 Hygromycine and 50 mg L−1 Kanamycine were added for the selection of transgenic cells. The induced calli were subcultured every 2 weeks on the same medium. The proliferated calli grown for 1 week after the second subculture at 25° or 15°C were excised and examined for GFP observation using a UV fluorescence microscope (Nikon, http://www.nikon.com) equipped with B-2A (450- to 490-nm excitation filter) for EGFP and UV-1A (360-nm excitation filter) for DAPI. In order to estimate the nucleus-to-cytoplasm GFP (NCG) ratio, GFP fluorescence intensities were measured by scanning of the microscopic images with imagej v1.43 (http://rsb.info.nih.gov./ij). An average of GFP fluorescence intensities at points with an equal distance within the nucleus or cytoplasm was used to estimate the NCG ratio. In each plot, 50 cells collected from at least three independent transgenic lines were randomly selected from the microscopic images and were analyzed to determine the NCG ratio.

Quantitative transient assay using GAL4 system

The plasmid pGBL has two operons, as shown in Figure 4a: one consists of the firefly luciferase gene (pGL4.12; Promega, http://www.promega.com) with TNOS sequences driven by the six tandem copies of the GAL4 binding site upstream of the TATA box; and the other, which is a reference to standardize differences in the efficiency of the bombardment delivery, carries the 35S promoter plus the Renilla luciferase gene (pRL-null; TOYO B-Net CO., LTD, http://www.toyo-b-net.co.jp/) with TNOS sequences. The segment containing the GAL4 binding sites and TATA box was adopted from pTA7001 (Aoyama et al., 1995) as a PCR segment with 5′-GACGTTGTAAAACGACGGCCAC-3′ and 5′-AGGCCAGCGTGTCCTCTCC-3′. pTA7001 was provided by Dr N.H. Chua (The Rockefeller University, New York, USA). In the two effector plasmids, pGATam3 and pGAGUS, GAL4 DNA-binding domain (amino acid positions 1–74) and activator of VP16 (positions 75–154) were inserted following a viral translation enhancer (Ω) present at the 3′ end of the 35S promoter sequence. The segments of GAL4 BD and activator were also amplified from pTA7001 with 5′-ATGAAGCTACTGTCTTCTATCGAAC-3′ and 5′-CCCACCGTACTCGTCAATTCC-3′. As shown in Figure 4, each of pGATam3 and pGAGUS contains Tam3 TPase gene or GUS gene, respectively, followed by the TNOS sequence.

The introduction of the plasmids into the flower petals (HAM5 line) was performed by the bombardment method. A mixture of pGBL (approximately 2.0 μg) and pGATam3 or pGAGUS (approximately 0.8 μg) was precipitated onto 1.2 mg of gold particles (1.6 μm in diameter) with ethanol, and 0.12 mg of the plasmid DNA-coated gold particles was delivered into intact tissue as described above. The bombarded buds were incubated for 16 h at 25° or 15°C. As the luciferase detection system is very sensitive, the pre-incubation of the bombarded buds at 25°C was not necessary. Firefly luciferase and Renilla luciferase activities were assayed using a Dual-luciferase Assay Kit (Promega), in accordance with the manufacturer’s instructions. Chemical luminescence was measured using a luminometer (Mithras, LB 940; Berthold Technology, http://www.berthold.com).

PEG transformation and microscopy observation

Protoplasts were isolated from young leaves of Antirrhinum plants (line HAM5) and tobacco BY-2 cells. Incubation of the protoplasts in isolation buffer (1.5% Cellulase Onozuka R10, 0.05% Pectolyase Y-23, 0.55 m mannitol) was performed for 3 h at a temperature at 15° or 25°C. For polyethylene glycol (PEG) transformation, purified protoplasts were washed in 0.55 m mannitol, 0.5% 2-(N-morpholine)-ethanesulphonic acid (MES), pH 5.6, and carrier salmon sperm DNA and plasmid DNA were added to 100 μl of protoplast suspension per tube. Subsequently, the same volume of 40% PEG 4000 was added. These protoplasts were washed once and incubated in W5 solution (154 mm NaCl, 125 mm CaCl2, 5 mm KCl and 5 mm glucose, pH 5.8) for 24 h at 25°C with 20 nm LMB until the observation. The protoplasts expressing GFP fusion proteins were observed using an UV fluorescence microscope (Olympus) equipped with the U-MNIBA filter set (470- to 490-nm excitation filter; 505-nm dichroic mirror; 515- to 550-nm barrier filter) for enhanced GFP.


We are especially grateful to Yoshio Sano (Laboratory of Plant Breeding) and Cathie Martin (Department of Cell and Developmental Biology, John Innes Center) for comments on the manuscript, and Yohsuke Takahashi (Graduate School of Science, Hiroshima University) for providing the plasmid pJ4-RSG-GFP. This work was supported by the Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (BRAIN) and Grant-in-Aid for Young Scientists from MEXT (no. 02J00614 to S.-N.H).