A 13-bp cis-regulatory element in the LTR promoter of the tobacco retrotransposon Tto1 is involved in responsiveness to tissue culture, wounding, methyl jasmonate and fungal elicitors

Authors


*For correspondence (fax +81 298 387408; e-mail hirohiko@abr.affrc.go.jp).

Summary

The tobacco Tto1 is one of the few active LTR-retrotransposons of plants, and its transposition is activated by tissue culture and is primarily regulated at the transcriptional level. The expression of Tto1 RNA can also be activated by various stresses, including viral infection, wounding, and treatment with jasmonate, a signal molecule of plant defence responses. It is shown here that the Tto1 LTR promoter is responsible for a high level of expression in cultured tissues of transgenic tobacco plants. We demonstrate that a 13-bp repeated motif (TGGTAGGTGAGAT) in the LTR functions as a cis-regulatory element, which confers the responsiveness to tissue culture, wounding and methyl jasmonate. Fungal elicitors also activate the promoter containing multiple copies of the 13-bp motif. Expression mediated by the 13-bp motif is activated markedly by okadaic acid and moderately by K252a, so that both phosphorylation and dephosphorylation of proteins are possibly involved in the signalling pathways. Interestingly, the 13-bp motif contains a conserved motif, Box L (also called AC-I or H-box like sequence) which has been shown to be involved in the expression of phenylpropanoid synthetic genes. Moreover, extended homologies are found between promoters of Tto1 and an asparagus defence gene, AoPR1, suggesting a possibility that the ancient insertion of an ancestral Tto1-related retrotransposon has provided some of the promoter/regulatory sequences, including the 13-bp motif-related sequence, of the AoPR1 gene. Based on the structural and functional similarity between the two promoters, a possible evolutionary role of the regulatory sequences of LTR-retrotransposons is discussed.

Introduction

Retrotransposons (or class I elements) are transposable elements that propagate via reverse transcription of RNA intermediates. In plants, diverse sequences of retrotransposons have been found in more than 100 species. Some retrotransposons are shown to have contributed to genome evolution by changing structures and expression patterns of genes (for a review, see Wessler et al. 1995). In maize and Vicia species, amplification of retrotransposons is in part responsible for the large genome sizes (Pearce et al. 1996; SanMiguel et al. 1996). Despite a wide and abundant distribution, however, most of the retrotransposon-related sequences in plants are presumed to be inactive fossils of ancient mobile elements, because of their defective structures. Tnt1 and Tnp2 of tobacco, and Bs1, stoner, B5, G, Magellan, Hopscotch and Zeon-1 of maize are the only retrotransposons responsible for recent mutations (reviewed in Grandbastien 1998). Although some plant retrotransposons have been reported to be transcriptionally active (Grandbastien 1998), only Tto1, Tnt1 and Tnp2 of tobacco and Tos17 of rice have been demonstrated to be mobile (Grandbastien et al. 1989; Hirochika et al. 1996; Hirochika 1993; Vaucheret et al. 1992). Even in the case of these active plant retrotransposons, most of the activity seems to be limited under stress conditions or in restricted tissues (Grandbastien 1998).

The tobacco Tto1 is a well characterised, autonomous LTR (long-terminal repeat)-retrotransposon (Hirochika & Otsuki 1995; Hirochika et al. 1996; Hirochika 1993; Takeda et al. 1998). Tto1 is almost inactive in unstressed vegetative tissues but can be activated by tissue culture. Therefore, propagation of Tto1 is considered to be a cause of tissue culture-derived genetic variability, called ‘somaclonal variation’. Transposition of Tto1 is primarily regulated at the transcriptional level, because it occurs concomitantly with a marked increase in levels of RNA. Recently, transcription of Tto1 was shown to be activated not only by tissue/cell culture but also by several defence-related stresses, such as viral infection, wounding, and treatment with jasmonate and methyl jasmonate (Hirochika 1993; Hirochika 1995; Takeda et al. 1998). Jasmonate and methyl jasmonate are involved in the activation of signal transduction pathways in response to wounding and pathogen attack in plants (Creelman & Mullet 1997; Wasternack & Parthier 1997). Activation of another plant LTR-retrotransposon, Tnt1, by stress has also been extensively studied. Expression of Tnt1 and/or Tnt1 LTR promoter-β-glucuronidase (GUS) fusion gene is not observed in tissues of normally grown tobacco plants except for roots, but induced by protoplasting, microbial elicitors, pathogen infection, wounding and some abiotic stresses (Grandbastien et al. 1997; Pouteau et al. 1991; and refs. therein). Although the biological meaning of activation of plant retrotransposons by stresses has not yet been fully established, the stress induced activity is consistent with the concept of ‘genomic stress’, proposed by McClintock (1984), based on the studies of class II elements, such as Ac/Ds and Spm of maize.

Promoters of retrotransposons are useful models for studying cis-regulatory elements involved in regulation of not only themselves but also cellular genes of host organisms, because of the following advantages. (1) Promoters of retrotransposons are usually compact in size, and it is easy to distinguish 5′-promoter ends from their flanking sequences. (2) Promoters of retrotransposons are highly mutable and selected to be adaptive in host genomes, so that regulatory sequences in enhancer regions are highly condensed and optimised in terms of their activities. For example, an enhancer region of the mouse VL30 LTR-retrotransposon contains binding sites for host-encoded regulatory proteins, such as AP-1, CREB, JUN, CarG, and NF-1 (Nilsson & Bohm 1994). (3) Cis-regulatory elements are often repeated in promoters of retrotransposons. The Drosophila copia retrotransposon contains repeated motifs similar to the SV40 core enhancer and repeated binding sites for homeoproteins (Cavarec & Heidmann 1993). The repeated sequence, BII box, in the tobacco Tnt1 also functions as an activator element (Casacuberta & Grandbastien 1993; Vernhettes et al. 1997).

Here, we report that the 13-bp repeated motif (TGGTAGGTGAGAT) in the Tto1 LTR functions as a cis-element involved in the response to tissue culture, wounding, methyl jasmonate, and some fungal elicitors. The 13-bp motif contains a Box L motif (also called AC-I or H-box like sequence), which has been shown to be involved in the expression of phenylpropanoid synthetic genes (Hatton et al. 1995; Loake et al. 1992; Logemann et al. 1995; Lois et al. 1989; Sablowski et al. 1994; Seki et al. 1997). In contrast to the previous motifs that require G-box for their functions in vivo, however, the 13-bp motif is sufficient to confer the stress inducibility. The role of the 13-bp motif is discussed in relation to the evolution of gene promoters, based on the structural and functional similarity between promoters of the Tto1 and AoPR1, an asparagus defence gene.

Results

Tto1 LTR promoter is responsible for the activation by tissue culture

To examine whether the Tto1 LTR promoter is responsible for the activation by tissue culture, we analysed the activity of the promoter, fused to a GUS reporter gene [LTR(– 199)-GUS], in leaf and callus tissues of transgenic tobacco. As shown in Fig. 1, high levels of expression of GUS activity were detected in callus. The expression levels varied from line to line, but some of them were comparable to the levels of expression driven by cauliflower mosaic virus (CaMV) 35S promoter [35S-GUS]. By contrast, expression of LTR(– 199)-GUS was not significant in leaf, as reported previously (Takeda et al. 1998). The variability of GUS activity in callus of LTR(– 199)-GUS transgenic lines may partly reflect that callus represents the mixture of cells with different physiological states, including competency of induction of gene expression. Analysis of 5′ deletion derivatives of the promoter (Fig. 1) revealed that activation by tissue culture is positively regulated via several cis-regulatory regions upstream of the TATA box. Deletion to – 96 caused significant reduction in GUS activities in callus, but the activity is still over 20 times higher than in leaves. Deletion to – 37 almost completely abolished the promoter activity, so that the region between – 96 and – 37 contains a minimal regulatory element. Similar results have been obtained by transient assay with the chloramphenicol acetyltransferase (CAT) reporter constructs in tobacco BY2 protoplasts (Hirochika et al. 1996), although protoplast preparation itself activates the Tto1 expression (Hirochika 1993).

Figure 1.

Activities of the Tto1 LTR promoter in callus tissues.

(Upper) Diagram of the Tto1 and the Tto1 LTR promoter. Locations of LTRs, coding regions for Gag (gag) and Pol (pol), the 13-bp and 15-bp directs repeats (closed box), and TATA box (open box) are shown. The intact [LTR(– 199)] and the various 5′-truncated promoter regions [LTR(– 164) to (– 37)] of Tto1 were fused to the GUS coding sequences. The numbers (– 199 to – 37) indicate the deletion end positions relative to the transcription start site (+ 1).

(Lower) GUS activities in leaves and callus derived from leaves of the indicated numbers of transgenic line (n) carrying the respective promoter-GUS fusion are plotted. Each symbol shows the value obtained with each independent transgenic line (in some cases, symbols are overlapped). Bars represent the average GUS activities.

The 13-bp repeated motif in LTR is a functional cis-regulatory element involved in the expression in protoplasts and callus

Tto1-LTR sequences, required for the high level expression in callus, contain the 15-bp and 13-bp direct repeats (Fig. 1). Functional significance of the 13-bp motif is also suggested by the deletion analysis (Fig. 1). Because transcriptional enhancers are often found as repeated sequences in promoters of retrotransposons (see Introduction), we examined whether the 15-bp and 13-bp repeats have regulatory functions, by transient assay in tobacco BY2 protoplasts. As shown in Fig. 2, the control LTR(– 37) promoter drives only low levels of expression of the CAT reporter [LTR(– 37)-CAT; Hirochika et al. (1996)], but insertion of the multiple copies of the 13-bp motif upstream of the basal promoter caused a significant increase in promoter activity [13AB-LTR(– 37)-CAT]. By contrast, multiple copies of the 15-bp motif could not enhance the promoter activity, significantly (data not shown). Three types of mutations in the 13-bp motif abolished or diminished the expression [13 CD-LTR(– 37)-CAT, 13EF-LTR(– 37)-CAT, and 13GH-LTR(– 37)-CAT], indicating that the promoter activity is dependent on the 13-bp motif. The 13-bp motif was also effective upstream of the CaMV 35S minimal promoter [35S(– 40)][13AB-35S(– 40)-CAT and 13EF-35S(– 40)-CAT]. These results indicate that the 13-bp motif functions as an enhancer element.

Figure 2.

The 13-bp motif functions as an enhancer element.

(a) Schematic structure of the promoter used for gain-of-function analysis. The wild-type 13-bp motif (13AB) was fused as nanomer (x 9) to the LTR basal promoter [LTR(– 37)] or the CaMV 35S minimal promoter [35S(– 40)]. The mutant 13-bp motifs (13 CD, 13EF, and 13GH) were fused as octamer (x 8) or nanomer (x 9) to the same promoter fragments. Nucleotide sequences of wild-type and three mutant variants of the 13-bp motif are shown.

(b) Promoter activity in tobacco protoplasts. Plasmids containing the indicated promoter-CAT constructs were introduced into protoplasts by electroporation. CAT activity in the transient expression assay is expressed relative to that of 13AB-LTR(– 37)-CAT (upper graph) or 13AB-35S(– 40)-CAT (lower graph) taken as 1. Average CAT activity with standard deviation obtained from three independent experiments is shown.

We next examined the effects of the 13-bp motif on the expression in callus, by using transgenic tobacco carrying the 13AB-LTR(– 37) promoter fused to the GUS reporter [Fig. 3; 13AB-LTR(– 37)-GUS]. High level expression of 13AB-LTR(– 37)-GUS was observed in callus, compared with the expression of LTR(– 37)-GUS, but not in leaf. The expression of 13AB-LTR(– 37)-GUS in callus from several transgenic lines was considerably strong, compared with that of 35S-GUS (Fig. 1). Mutational analysis showed that the expression in callus is dependent on the 13-bp motif [Fig. 3; 13 CD-LTR(– 37)-GUS to 13GH-LTR(– 37)-GUS]. The 13-bp dependent, callus-specific expression was also observed with the constructs based on the 35S(– 40) promoter [35S(– 40)-GUS, 13AB-35S(– 40)-GUS and 13EF-35S(– 40)-GUS], thus clearly indicating that the 13-bp motif confers the responsiveness to tissue culture.

Figure 3.

The 13-bp motif confers the responsiveness to tissue culture. The promoter fragments, composed of the wild-type (13AB) or the mutated (13 CD to 13GH) 13-bp motifs and the LTR(– 37) or the 35S(– 40) minimal promoter (as shown in Fig. 2a), were fused to the GUS coding sequences. GUS activities in leaves and callus derived from leaves of the indicated numbers of transgenic line (n) carrying the respective promoter-GUS fusion gene are plotted. Each symbol shows the value obtained with each independent transgenic line (in some cases, symbols are overlapped). Bars represent the average GUS activities.

The 13-bp motif confers responsiveness to wounding and methyl jasmonate

Previous studies showed that the Tto1 promoter contains cis-regulatory regions responsive to wounding and methyl jasmonate (Takeda et al. 1998). Because the LTR(– 132)-promoter (Fig. 1) can also be activated by these stresses, we examined whether the 13-bp motif is involved in the activation. When the leaves of 13AB-LTR(– 37)-GUS transgenic lines were excised, cut into segments, and incubated for 1–2 days, expression of GUS activity was induced (Fig. 4a). The induction began to be detected within 8 h after cutting (data not shown). In five of seven transgenic lines examined, additional wounding by stabbing of cut segments resulted in increased GUS activity (Fig. 4a; Cut + Wn, induction ratios are also given in the legend). In addition, the activity of GUS after wounding treatment was localised mainly at the vicinity of the injured sites (Fig. 4b; both the cut site and the stabbed sites), suggesting that the response is actually dependent on wounds. These expression patterns of the 13AB-LTR(– 37)-GUS after cutting and stabbing wounding are quite similar to those of Tto1 RNA or LTR(– 199)-GUS (Takeda et al. 1998). Methyl jasmonate also activate the expression of the 13AB-LTR(– 37)-GUS (Fig. 5). The activation by wounding and methyl jasmonate was confirmed by RNA analysis (data not shown). By contrast, neither the wound-or methyl jasmonate-induced expression was observed with LTR(– 37)-GUS. Three types of mutations (13 CD, 13EF, and 13GH) resulted in loss or decreased levels of the responsiveness to wounding and methyl jasmonate, although 13GH confers an increase in constitutive expression in leaves (Fig. 5). The 13-bp dependent expression by wounding and methyl jasmonate was further con- firmed with 35S(– 40) promoter [13AB-35S(– 40), and 13EF-35S(– 40)]. Thus, it is concluded that the 13-bp motif also confers responsiveness to wounding and methyl jasmonate.

Figure 4.

Activation of the 13AB-LTR(– 37)-GUS fusion gene by wounding.

(a) Leaf segments excised from transgenic plants were incubated on 0.05% MES buffer for indicated periods with (+ Wn) or without (Cut) stab wounding. GUS activities in the segments of the indicated numbers of transgenic line (n) are plotted. Each symbol shows the value obtained with each independent transgenic line. Bars represent the average GUS activities. Ratios of induction with stabbing to without stabbing, [(Cut + Wn, 2 days) – (0 day)]/[(Cut, 2 days) – (0 day)], from the seven lines are 4.0, 3.3, 1.8, 0.9, 2.0, 2.5, and 0.5. Similar results were obtained by the experiments with leaf segments incubated on water.

(b) Localisation of GUS activities in the wounded leaf. A detached leaf of transgenic tobacco plant was stabbed with forceps only in the right-half side of the midrib. The leaf was stained with X-gluc after incubation for 2 days. Similar results were obtained from several independent transgenic lines.

Figure 5.

The 13-bp motif confers the responsiveness to wounding and MJA.

The wild-type (13AB) and the mutated (13 CD to 13GH) 13-bp motifs were tested for wound and MJA inducibility in leaves of transgenic tobacco plants. (Left) GUS activities in leaves before (not treated) and after treatment with cutting, stabbing and incubation for 48 h (wounded). (Right) GUS activities in leaf segments after incubation in the presence (MJA) or absence (control) of 50 μM MJA for 24 h. Each symbol shows the value obtained with each independent transgenic line (in some cases, symbols are overlapped). Bars represent the average GUS activities obtained from the indicated numbers of transgenic line (n).

Expression of Tto1 and 13AB-LTR(– 37)-GUS is activated by fungal elicitors

Protoplast preparation from cultured cells or leaves has been shown to activate the expression of Tnt1 and Tto1 (Pouteau et al. 1991; Hirochika 1993), raising the question of whether the same or related factors are involved in the activation. A cause of induction of Tnt1 expression is an elicitor activity in fungal extracts, used as a protoplasting enzyme (Onozuka R10 cellulase)(Pouteau et al. 1991). Tnt1 can also be activated by other microbial elicitors of plant defence responses (Grandbastien et al. 1997 and references therein). So, we examined whether fungal elicitors activate the expression of Tto1 and 13AB-LTR(– 37)-GUS in leaves by RNA blot analysis. Defence responses induced by elicitors were verified with the expression of endogenous wound-and elicitor-inducible genes for PAL (Pellegrini et al. 1994).

As shown in Fig. 6, Onozuka R10 extracts clearly induced the expression of Tto1 and 13AB-LTR(– 37)-GUS, as well as PAL genes. Therefore, the fungal extracts might be a cause of activation of these genes in protoplasts, which are prepared with higher concentrations of Onozuka extracts (10 mg ml–1; Hirochika 1993). Heat-treatment of the extracts decreased the expression of Tto1 and 13AB-LTR(– 37)-GUS, but not completely, and still allowed to induce the expression of PAL genes significantly (Onozuka R10 autoclaved). Thus, elicitor activities in Onozuka R10 extracts consist of heat-sensitive and heat-resistant components. The effects of the elicitor activities on the expression of 13AB-LTR(– 37)-GUS are most likely dependent on the 13-bp motif, because the expression in protoplasts is abolished or diminished by 13 CD, 13EF and 13GH mutations (Fig. 2).

Figure 6.

Expression of 13AB-LTR(– 37)-GUS after treatment with fungal elicitors.

Leaves of transgenic tobacco plants carrying 13AB-LTR(– 37)-GUS were treated with MES buffer in the absence [(MES)] or presence of 1 mg ml–1 Onozuka R10 extracts (Onozuka R10), 1 mg ml–1 heat-treated Onozuka R10 extracts (Onozuka R10 autoclaved), 100 μM chitin oligomer, or 2.5 μg/g tissue of xylanase for 8 h. Total RNA (20 μg) from leaf tissues was analysed by RNA blot analysis with the indicated DNA probes. Levels of 25S rRNA on the blotted membrane are shown by staining with methylene blue (rRNA). NT indicates untreated leaf tissues.

Several commercial enzymes for protoplast isolation, including cellulase Onozuka, have been shown to contain activity of xylanase (Fuchs et al. 1989; Nagata et al. 1981), which degrades hemicellulose components of plant cell walls. Xylanase has an elicitor activity to induce a range of defence responses in tobacco, including ethylene biosynthesis (Fuchs et al. 1989) and pathogenesis- related (PR) protein synthesis (Lotan & Fluhr 1990; Raz & Fluhr 1993). Therefore, xylanase may account for the elicitor activity in Onozuka R10 extracts at least in part. Consistent with this, the expression of Tto1, 13AB-LTR(– 37)-GUS and PAL genes was strongly induced by xylanase (Fig. 6).

Fungal cell wall components, such as chitin fragments, are also known to induce defence responses in plants (Ryan & Farmer 1991). The expression of Tto1, 13AB-LTR(– 37)-GUS, and PAL genes, was moderately in- duced by chitin oligomers (Fig. 6). Taken to- gether, these results indicate that expression of Tto1 and 13AB-LTR(– 37)-GUS is associated with defence responses activated by fungal elicitors.

The 13AB-LTR(– 37) promoter is strongly activated by a protein phosphatase inhibitor

It has been reported that protein kinases are involved in the signalling of various plant defence responses, and that exogenous supply of inhibitors of protein phosphatases, such as okadaic acid, can mimic such responses (Gianfagna & Lawton 1995; Raz & Fluhr 1993). As shown in Fig. 7, the expression of 13AB-LTR(– 37)-GUS in cut leaf segments was markedly enhanced by treatment with okadaic acid (OKA), compared with the control treatment (DMSO). Likewise, okadaic acid activated the LTR(– 199) and the 13AB-35S(– 40) promoters, but not the LTR(– 37) and the 13EF-LTR(– 37) promoters (data not shown). These results suggest that protein kinases are involved in the signal transduction pathways leading to gene activation via the 13-bp motif. However, K252a, a protein kinase inhibitor, did not repress but moderately enhanced the expression of 13AB-LTR(– 37)-GUS (Fig. 7), suggesting that positive regulation by protein phosphatase(s) is also involved. The results are similar to those observed with the expression of wound-inducible JR3 gene of Arabidopsis (Rojo et al. 1998). In Arabidopsis, at least two wound signal transduction pathways are suggested; one is a jasmonate-dependent pathway that is negatively regulated by a protein kinase, and another is a jasmonate-independent pathway that is positively regulated by a protein kinase. Expression of JR3 gene, possibly regulated via both pathways, is activated markedly by okadaic acid and moderately by staurosporine, a protein kinase inhibitor. Likewise, two or more independent signalling pathways may be involved in the 13-bp mediated gene activation in tobacco.

Figure 7.

Activation of 13AB-LTR(– 37)-GUS by inhibitors of protein kinases and protein phosphatases.

Cut leaf segments, prepared from transgenic tobacco plants carrying 13AB-LTR(– 37)-GUS, were treated with 0.05% MES buffer containing 0.5% DMSO in the absence [(DMSO)] or presence of 5 μm K252a or 0.5 μm okadaic acid (OKA) for 16 h. GUS activities in the segments are plotted. Each symbol shows the value obtained with each independent transgenic line. Bars represent the average GUS activities obtained from three independent transgenic lines. NT indicates untreated segments.

Discussion

The 13-bp motif is a cis-regulatory element, which confers the responsiveness to tissue culture, wounding, and methyl jasmonate

It is shown here that the Tto1 LTR promoter is responsible for the activation by tissue culture and that several enhancer regions are involved in the activation (Fig. 1). By gain-of-function analysis, the 13-bp motif was shown to confer the responsiveness to tissue culture, wounding and methyl jasmonate (Figs 2–5). The multimerized 13-bp motif was functional by fusion with either the LTR(– 37) promoter or the heterologous 35S(– 40) promoter, but the levels of GUS activity driven by the 13AB-35S(– 40) promoter were much lower than those driven by the 13AB-LTR(– 37) promoter (Figs 3 and 5). Therefore, the minimal promoter unit or downstream region of the Tto1 LTR, including untranslated leader sequence, might have a function to optimise the expression by the 13-bp repeat. The three types of mutations in the 13-bp motif have almost similar effects on the expression induced by the stresses examined, suggesting that the same or related transcription factors are involved in the activation via the 13-bp motif. Fungal elicitors also activated the expression of 13AB-LTR(– 37)-GUS, nearly correlated to the expression of Tto1 and tobacco PAL genes (Fig. 6). Thus, the 13AB-LTR(– 37) promoter can be a useful marker to pick out the specific signalling pathway(s) leading to gene expression by limiting the potential combination of the cis-elements and trans-factors.

By transient assays, a significant enhancer activity of the 15-bp repeated motif was not detected. However, we cannot exclude that the 15-bp motif may have some functions in regulation of Tto1 under other stress conditions. The repeated motif may be insufficient for the putative function, since it can be extended into a 21-bp imperfect repeat, which may be a functional element [Fig. 8(b) 15-bp motif (extended)].

Figure 8.

The 13-bp motif contains the cis-regulatory sequences conserved among defence gene promoters.

(a) The complementary sequence of the 13-bp motif of Tto1 is compared with the motifs in the indicated gene promoters (athis study; bLois et al. 1989; cHauffe et al. 1991; dWarner et al. 1993; eLoake et al. 1992).

(b) Sequence comparison of the promoter regions of AoPR1 and Tto1. Locations of the promoter fragments are indicated by positions relative to the transcriptional start sites (Hirochika 1993; Warner et al. 1993). TATA box sequences (TATA) are indicated with italics and by arrows. The direct repeats in AoPR1[repeat (AoPR1)], the 13-bp motifs, the 15-bp motifs, and the 21-bp imperfect repeats [15-bp motif (extended)] in Tto1 are indicated by distinct arrows. Identical nucleotides are indicated by capital letters and by asterisks. Gaps are indicated by dotted lines.

The 13-bp motif shows homologies to motifs involved in the activation of defence genes

The 13-bp motif is not similar to the previous jasmonate-responsive elements (Creelman & Mullet 1997; Wasternack & Parthier 1997), but we found that the complementary sequence of the 13-bp motif contains box L (also called AC-I motif) and the core sequence of H-box element (Fig. 8a), which are highly conserved among phenylpropanoid biosynthetic gene promoters (Hatton et al. 1995; Loake et al. 1992; Logemann et al. 1995; Lois et al. 1989; Sablowski et al. 1994; Seki et al. 1997). Extensive studies have revealed that these conserved cis-elements are involved in the stress-responsive and/or tissue-specific expression. However, mechanisms that regulate the expression of genes through the previous boxes might be different from those which regulate the expression of Tto1 through the 13-bp motif. Tandem copies of the PAL-1 box L could not confer the responsiveness to UV light and to an elicitor derived from Phytophthora megasperma f.sp. glycinea in parsley protoplasts (Logemann et al. 1995). In some promoters, box L/AC-I or H-box motifs function in combination with the adjacent G-box (CACGTG) or G-box like elements as tissue-specific or stress responsive regulatory units (Faktor et al. 1997; Loake et al. 1992; Sablowski et al. 1994; Seki et al. 1997). By contrast, the repeated 13-bp motif sufficiently confers the responsiveness to stresses, and no G-box is found in the Tto1 promoter.

Several factors binding to L/AC-I/H-box motif have been cloned. Among them, Myb factors were shown to mediate the tissue specific expression (Grotewold et al. 1994; Sablowski et al. 1994). A soybean factor binding to both G-box and H-box, G/HBF-1, was activated in DNA binding activity after phosphorylation by a specific kinase stimulated in elicited cells with glutathione or an avirulent pathogen Pseudomonas syringae pv. glycinea (Dröge-Laser et al. 1997). A parsley Myb-related factor binding to box L, PcMYB1, is a candidate for regulators of light-mediated gene activation, but suggested to function in conjunction with other regulator(s) (Feldbugge et al. 1997). Recently, several tobacco Myb factors binding to the 13-bp motif have been cloned (K. Sugimoto, S. Takeda, and H. Hirochika, unpublished), and functions of the factors are currently investigated.

A possible evolutionary role of the regulatory sequences of retrotransposons

Transposable elements have been suggested to contribute to the evolution of genes, e.g. by providing cis-regulatory elements leading to changes in expression patterns of genes (Kidwell & Lisch 1997; McDonald 1990; Wessler et al. 1995). In plant genomes, retrotransposon-derived sequences have frequently been found in promoter regions of genes (White et al. 1994). Interestingly, we found that the promoter of an asparagus defence gene, AoPR1, contains several regions homologous to the complementary sequence of the Tto1 LTR promoter, including the 13-bp/H-box like motif (Fig. 8a,b; Warner et al. 1993). Striking conserved sequences are located within a 13-bp motif and within a 21-bp imperfect repeat of Tto1 [Fig. 8(b) 13-bp motif and 15-bp motif (extended)]. Moreover, the AoPR1 promoter contains a sequence (–288cCTCTGAgACgA–277) similar to the 3′-end of the initiator methionine tRNA (5′-gCTCTGAtACcA-3′), whose complementary sequence corresponds to a primer binding site for reverse transcription. These findings suggest a possibility that an ancient insertion of an ancestral Tto1-related retrotransposon has provided some of the promoter sequences of the AoPR1 gene. Although the conserved sequences between the two promoters are restricted, this is consistent with the hypothesis that only adaptive regulatory sequences, provided by transposable elements, are preserved after selection over a long span of evolutionary time (McDonald 1990).

The AoPR1 transciption start site is located inside of the Tto1-related sequences (Fig. 8b). Notably, AoPR1 promoter can be activated by stresses, such as tissue culture, wounding (in adjacent regions from wounded sites), pathogen infection and salicylic acid (Firek et al. 1993; Warner et al. 1993; Warner et al. 1994), as can the Tto1 promoter be. In addition, activation of both genes by salicylic acid requires cutting/wounding stimuli (Hirochika 1995; Mur et al. 1996; H. Hirochika, unpublished results). Considering these similarities, it is speculated that the Tto1-related remnants have provided a useful function to the AoPR1 gene. Although definitive evidences are needed to conclude, if it was the case, that would be the first example of a striking correlation between the ‘creation’ of a cellular regulatory sequence, by insertion of a LTR-retrotransposon, and the current ‘natural’ function of the corresponding cellular genes. In this regard, our findings would support the hypothesis that ‘LTR-retrotransposons constitute drive mechanisms for the evolution of enhancers, and subsequent distribution of these enhancers throughout eukayotic genomes may play a significant role in eukaryotic regulatory evolution’ (McDonald et al. 1997).

The similarity between the two promoters also suggests that the asparagus genome have contained the retrotransposons that have a common origin with Tto1. It further raises a possibility that this lineage of elements once had been a ubiquitous component of plant genomes, unless horizontal transfer between dicots and monocots had occurred. At present, among active plant LTR-retrotransposons, only Tto1, Tnt1 and Tto5 of tobacco have been reported to be activated by defence-related stresses (Grandbastien 1998). Tnt1 is most similar to Tto1, so that Tnt1 may belong to the same lineage as that Tto1 belongs to. It should be noted that the Tnt1 LTR also contains H-box like sequences in the functional enhancer regions (Grandbastien 1998; Vernhettes et al. 1997; and references therein). Further characterisation of the regulatory features of retrotransposons and genes, together with progresses in genome analysis, will help to illustrate to what extent retrotransposons have contributed to the regulatory evolution in plants.

Experimental procedures

Plant materials and treatment of plant tissues

Preparation of transgenic tobacco (Nicotiana tabacum cv. SR1) and treatment of leaves for induction by wounding and methyl jasmonate were performed as described (Takeda et al. 1998). For callus induction, leaf segments from in vitro plants were incubated for 17 days on medium, containing Murashige-Skoog salts (Dainippon Pharmaceutical Co., Osaka, Japan), 3% sucrose, 2 mg l–1 1-naphthaleneacetic acid (NAA), 0.25 mg l–1 6-benzylaminopurine (BA) and 1 mg l–1 thiamine-HCl. Okadaic acid (Wako Pure Chemical Industries, Osaka, Japan) and K252a (Calbiochem, San Diego, CA) were dissolved in dimethyl sulfoxide (DMSO) at 1 mm for storage, and diluted in 2-[N-morpholino] ethanesulfonic acid (MES)-KOH buffer (pH 5.7). Leaf segments were floated on the buffered solutions containing these inhibitors and 0.5% DMSO for 16 h at room temperature.

For treatment with elicitors, progenies of transgenic plants were grown in a green house. Detached leaf-petioles were fed with 0.05% MES-KOH buffer (pH 5.7) in the presence or absence of Cellulase Onozuka R10 extracts from Trichoderma viride (Yakult, Pharmaceutical Co. LTd, Hyogo, Japan) or chitin oligosaccaride (5–6 mers; Higher chitin-oligosaccharide CH, Seikagaku, Tokyo, Japan) for 8 h at room temperature. Non-denatured Onozuka R10 extracts were dissolved in water and filtrated with 0.2 μM cellulose nitrate microporous membranes, and heat-denatured Onozuka R10 extracts were prepared by autoclaving at 120°C for 20 min before filtration. Xylanase purified from Trichoderma viride (Sigma, St Louis, MO, USA) was applied as a hanging drop (2.5 μg g–1 tissue) to freshly cut sites of detached leaf-petioles. After the solution was absorbed, the leaf-petioles were fed with 0.05% MES-KOH buffer (pH 5.7). After treatment, leaf portions without mid ribs were used for RNA analysis.

Protoplasts from tobacco BY2 cells were prepared as described previously (Hirochika 1993; Nagata et al. 1981).

Plasmid construction

For plant transformation, GUS fusion genes were cloned into pBI based vectors (Clonetech). pBILTRGUS-1, containing LTR(– 199)-GUS [previously designated LTR(1–1):GUS,] was constructed previously (Takeda et al. 1998). The deletion derivatives were constructed, using promoter fragments prepared from pLTRCAT-2, pLTRCAT-3, pLTRCAT-4, and pLTRCAT-5, respectively (Hirochika et al. 1996). The HindIII-SpeI fragment of each truncated promoter and a SpeI-HincII fragment derived from pSKTto1 (Hirochika et al. 1996), carrying the 5′-non-coding region and the region coding for N-terminal nine amino acids, were cloned between HindIII and SmaI sites of pBI101.3. The resulting plasmids pBILTRGUS-2, pBILTRGUS-3, pBILTRGUS-4, and pBILTRGUS-5 contain LTR(– 164)-GUS, LTR(– 132)-GUS, LTR(– 96)-GUS, and LTR(– 37)-GUS, respectively. pBI121 was used for CaMV 35S-GUS construct.

The promoter fragment of the minimal CaMV 35S promoter (– 40 to + 5) was prepared by polymerase chain reaction (PCR) amplification from p35SCAT (Hirochika et al. 1996). The amplified fragment was blunt-ended by filling in with the Klenow fragment, digested with XbaI, and cloned between HincII and XbaI sites of p35SCAT, resulting in p35S(– 40)-CAT.

The multimerized 13-bp promoter-CAT constructs were prepared as follows. Complementary oligonuculeotides (13AB; 5′-ctagaTGGTAGGTGAGATa-3′ and 5′-ctagtATCTCACCTACCAt-3′, 13 CD; 5′-ctagaGTTTAGGTGAGATa-3′ and 5′-ctagtATCTCACCTAAACt-3′, 13EF; 5′-ctagaTGGTCTTTGAGATa-3′ and 5′-ctagtATCTCAAAGACCAt-3′, 13GH; 5′-ctagaTGGTAGGTTCTATa-3′ and 5′-ctagtATAGAACCTACCAt-3′) were annealed and cloned between XbaI and SpeI site of a pBluescriptSK + vector (Stratagene), as a unit of two or three tandem repeats. Further concatenated XbaI-SpeI fragment was cloned into the XbaI site of pLTRCAT-5 (containing LTR(– 37) -CAT; Hirochika et al. 1996), resulting in p13AB-LTR(– 37)-CAT, p13 CD-LTR(– 37)-CAT, p13EF-LTR(– 37)-CAT, or p13GH-LTR(– 37)-CAT. The XbaI-SpeI fragment, containing 13AB (9mer) or 13EF (8mer), was also cloned into the HindIII site (filled in with the Klenow fragment) of p35S(– 40)-CAT, resulting in p13AB-35S(– 40)-CAT or p13EF-35S(– 40)-CAT.

The multimerized 13-bp promoter-GUS constructs were prepared as follows. The HindIII-SpeI promoter fragment of pBILTRGUS-1 was replaced with the HindIII-SpeI promoter fragments of p13AB-LTR(– 37)-CAT, p13 CD-LTR(– 37)-CAT, p13EF-LTR(– 37)-CAT, or p13GH-LTR(– 37)-CAT, to produce pBI13AB-LTR(– 37)-GUS, pBI13 CD-LTR(– 37)-GUS, pBI13EF-LTR(– 37)-GUS, and pBI13GH-LTR(– 37)-GUS, respectively. The HindIII-XbaI promoter fragment of p35S(– 40)-CAT was cloned between HindIII and XbaI sites of pBI101.3, resulting in pBI35S(– 40)-GUS. The XbaI promoter fragment of p13AB-35S(– 40)-CAT or p13EF-35S(– 40)-CAT, was cloned into XbaI site of pBI101.3, resulting in pBI13AB-35S(– 40)-GUS or pBI13EF-35S(– 40)-GUS.

Transient expression of CAT fusion genes and CAT assay

Introduction of various CAT constructs into BY2 protoplasts by electroporation, extraction of proteins, and assay of CAT activity were performed as described (Hirochika et al. 1996).

GUS staining and assay of GUS activity

GUS staining, extraction of proteins, and fluorometric assay of GUS activity were performed as described (Jefferson et al. 1987).

RNA blot hybridisation

Preparation of total RNA and RNA bolt hybridisation with 32P-labelled DNA probes for Tto1, GUS and PAL genes were performed as described (Hirochika et al. 1996). The Eco105I-SacI fragment of pBI221 was used for GUS specific DNA probes. A DNA fragment for PAL specific probes was prepared by PCR amplification from tobacco genomic DNA, using specific primers (5′-tgggaaatggcagctgaatc-3′ and 5′-aagagcaccaccattcttgg-3′). The amplified DNA fragment, corresponding to the tobacco PAL.E. gene (Pellegrini et al. 1994), was cloned in a plasmid and verified by DNA sequencing.

Acknowledgements

We would like to thank Dr R.A. Martienssen for helpful discussions. This work was supported by a project grant from the Ministry of Agriculture, Forestry, and Fisheries of Japan, a grant-in-aid for scientific research on priority areas from the Ministry of Education, Science, and Culture of Japan (the Molecular Basis of Flexible Organ Plans in Plants, no. 06278103), and an enhancement of Center-of-Excellence, special coordination funds for promoting science and technology in Japan.

Ancillary