SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

ACGT-containing ABA response elements (ABREs) have been functionally identified in the promoters of various genes. In addition, single copies of ABRE have been found to require a cis-acting, coupling element to achieve ABA induction. A coupling element 3 (CE3) sequence, originally identified as such in the barley HVA1 promoter, is found approximately 30 bp downstream of motif A (ACGT-containing ABRE) in the promoter of the Osem gene. The relationship between these two elements was further defined by linker-scan analyses of a 55 bp fragment of the Osem promoter, which is sufficient for ABA-responsiveness and VP1 activation. The analyses revealed that both motif A and CE3 sequence were required not only for ABA-responsiveness but also for VP1 activation. Since the sequences of motif A and CE3 were found to be similar, motif-exchange experiments were carried out. The experiments demonstrated that motif A and CE3 were interchangeable by each other with respect to both ABA and VP1 regulation. In addition, both sequences were shown to be recognized by a VP1-interacting, ABA-responsive bZIP factor TRAB1. These results indicate that ACGT-containing ABREs and CE3 are functionally equivalent cis-acting elements. Furthermore, TRAB1 was shown to bind two other non-ACGT ABREs. Based on these results, all these ABREs including CE3 are proposed to be categorized into a single class of cis-acting elements.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Since most promoters of ABA-inducible genes contain ACGTGGC motifs within 300 bp upstream of the transcription start sites, the motif was predicted to be an ABA response element or ABRE (Michel et al. 1993; Mundy et al. 1990; Yamaguchi-Shinozaki et al. 1989; reviewed by Busk & Pages 1998; Giraudat et al. 1994). Subsequent functional analyses of ABA-inducible promoters such as the wheat Em and rice Rab16A have led to the conclusion that the ACGTGGC motif can actually function as an ABRE. This conclusion has originated from two lines of evidence (Guiltinan et al. 1990; Skriver et al. 1991). First, it was demonstrated that a 75 bp fragment of the wheat Em promoter was able to confer ABA-responsiveness to the CaMV 35S promoter truncated at –90, and that a base-change mutation of the motif led to a great reduction of ABA-responsiveness (Guiltinan et al. 1990). Second, six copies of a short oligonucleotide containing this motif from Rab16A were found to exhibit ABA-responsiveness when they were fused to the CaMV 35S minimal promoter (Skriver et al. 1991). Later, the function of the ACGTGGC motif as an ABRE was also demonstrated in other ABA-inducible promoters (Busk & Pages 1997; Busk et al. 1997; Hattori et al. 1995; Ono et al. 1996; Shen & Ho 1995; Shen et al. 1996; Vasil et al. 1995). Since these ABREs contain a tetra mucleotide, ACGT, which is a typical core sequence of plant bZIP proteins (Izawa et al. 1993; Meshi & Iwabuchi 1995; Williams et al. 1992), we use the terminology of ‘ACGT-containing ABRE’ for this type of ABRE.

Although functional importance of ACGT-containing ABREs has been clearly demonstrated in ABA-regulated transcription, two puzzles remain. First, multiple but not single copies of ACGT-containing ABREs can confer ABA-responsivenenss to a heterologous minimal promoter (Skriver et al. 1991; Vasil et al. 1995). Second, sequences similar or identical to the ACGT-containing ABRE have been identified as cis-acting elements in the promoters of various genes which are regulated by other signals. The effectors include the cell-cycle (Nakayama et al. 1989), white light (Giuliano et al. 1988; Kao et al. 1996), UV-light (Schulze-Lefert et al. 1989), auxin (Liu et al. 1994; Ulmasov et al. 1995) and jasmoic acid (Mason et al. 1993). The presence of a second sequence element in the G-box vicinity has been demonstrated to play an important role in specifying the light-responsiveness of the parsley CHS promoter (Block et al. 1990). A second cis-acting element (TGCCACCGG) named CE1 (for coupling element 1) was identified by Shen & Ho (1995), who conducted a linker-scan analysis of a 44 bp, ABA-responsive fragment from the barley HVA22 promoter. In this fragment, CE1 was found to function with a single copy of an ACGT-containing ABRE, together constituting a minimal ABA-response complex (ABRC). Similar analyses with a different barley promoter (HVA1) have identified another coupling element CE3 (ACGCGTGTC) distinct from CE1 (Shen et al. 1996). In other instances, multiple ACGT-containing ABREs and other elements are clustered and appear to function in a complex (Busk & Pages 1997; Busk et al. 1997).

It has been postulated that ABA-responsive transcription in seed tissues is specified by the seed-specific transcription factor VP1 in maize (Hattori et al. 1992; McCarty et al. 1991) and the orthologous ABI3 in Arabidopsis (Giraudat et al. 1992). This hypothesis is strongly supported by the observation that the ABA responsiveness of otherwise seed-specific genes, such as AtEm1, can be extended to vegetative tissues in Arabidopsis when ABI3 is ectopically expressed with a CaMV 35S promoter (Parcy et al. 1994). Homologues or orthologues of VP1/ABI3 have been cloned from rice (OSVP1; Hattori et al. 1994), common bean (PvAlf; Bobb et al. 1995) and several other plant species. VP1 has been shown to transactivate target genes through at least two types of promoter element. For the anthocyanin regulatory gene, C1, data indicate that VP1 activation occurs essentially through an Sph element (Hattori et al. 1992). Sph element specific binding activity of the C-terminal conserved domain (B3 domain) of VP1 has recently been demonstrated by Suzuki et al. (1997). In contrast, VP1 regulation of promoters in LEA-type ABA-regulated genes is quite different from that of the C1 promoter (Carson et al. 1997). We have shown that a base-change mutation in the ACGT-containing ABRE, designated motif A in the Osem promoter, leads to a striking reduction of not only ABA-responsiveness but also VP1 activation (Hattori et al. 1995). Vasil et al. (1995) have also demonstrated that the activation of the Em promoter can be achieved through ACGT-containing ABREs by both loss-of-function and gain-of-function type experiments. They have shown that tetramers of the ACGT-containing ABREs (Em1a or Em1b) as well as of the G-box sequence from the parsley CHS promoter can confer VP1 activation to a 35S minimal promoter (Vasil et al. 1995). Although VP1 has been clearly shown to act through ACGT-containing ABREs in ABA-regulated genes, no direct binding activity to these elements has been detected. In addition, several cloned bZIP proteins, which include wheat EmBP-1 (Guiltinan et al. 1990) and tobacco TAF-1 (Oeda et al. 1991) and rice OSBZ8 (Nakagawa et al. 1996b) have been demonstrated to bind to ACGT-containing ABREs. From these observations, it has been postulated that VP1 functions by interacting with a factor that directly binds to ABREs (Hattori et al. 1995; Vasil et al. 1995). This hypothesis has been substantiated by recent cloning of a VP1-interacting ABRE-binding bZIP protein, TRAB1 (Hobo et al. submitted). In addition to its specific binding to ACGT-continaing ABREs and interaction with VP1, TRAB1, when expressed as a fusion protein with GAL4 DNA binding domain, can confer ABA-inducibility to a chimeric promoter with GAL4 binding sites. Thus, TRAB1 is proposed to be a true ABRE-binding factor that can mediate an ABA-signal.

In the present study, we conducted linker-scan analyses of a short Osem promoter fragment which contained a single copy of an ACGT-containing ABRE and a CE3-like sequence which was sufficient to confer both ABA and VP1-responsiveness to a heterologous minimal promoter. Data demonstrate that the CE3-like sequence is important not only for ABA-responsiveness but also for regulation by VP1. Furthermore, we have recognized a sequence similarity between the ACGT-containing ABREs and the CE3 sequence, and show here that the two elements share a functional equivalence and can be bound by TRAB1. Based on these results, we propose that two copies of G-box-related ABRE/CE3 sequences can constitute a VP1-regulated ABA-response complex. In addition, we show that TRAB1 can bind to other non-ACGT ABREs, indicating that ACGT-containing and non-ACGT ABREs and CE3 are the ABA-responsive regulatory sequences of essentially the same nature.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

A chimeric promoter consisting of a 55 bp fragment of the Osem promoter and the CaMV 35S minimal promoter responds to ABA and can be activated by VP1

We have previously shown, via transient expression assays in a rice protoplast system, that the promoter region of Osem up to –176 is sufficient to confer induction by exogenous ABA and activation by overexpression of OSVP1 (Hattori et al. 1995). Furthermore, we have determined that a ACGT-containing sequence (motif A) located at –171 has a critical role in both responses. Moreover, we have also identified a 17 bp sequence (‘region 1’) located just downstream of this motif A that appears to contain an important element(s) for the regulation by ABA and VP1 (Hattori et al. 1995). Deletion of region 1 greatly diminishes both responses. We subsequently identified a CE3 sequence just downstream of region 1 (at –142; Fig. 1a). Thus, we sought to test whether the CE3 sequence in the Osem promoter functions as a coupling element for motif A and whether it plays any role in VP1 regulation.

image

Figure 1. ABA induction and VP1 activation of a chimeric promoter consisting of the 101 or 55 bp fragments of the Osem promoter and the CaMV 35S minimal promoter (CaMV 35S TATA).

(a) Nucleotide sequences of the Osem and HVA1 promoter fragments that contain an ACGT-containing ABRE (boxed) and a CE3 sequence (circled). Region 1 of the Osem promoter is underlined.

(b–d) Transient assays with rice protoplasts. The constructs schematically illustrated in (b) were electroporated with the ubiquitin promoter-luciferase plasmid (an internal control) into rice protoplasts and tested for ABA induction (c) and VP1 activation (d). After incubation of the protoplasts for 40 h, GUS and luciferase (LUC) activities were determined. Values represent the means (± SEM) of GUS/LUC activity ratios which resulted from three independent electroporations. Data shown in each panel are those obtained in an experiment using a single preparation of protoplasts. X indicates fold induction (c) or activation (d).

Download figure to PowerPoint

To answer these questions, we first made two constructs to define a short cis-acting element complex that can minimally confer both ABA and VP1 responsiveness to a heterologous minimal promoter. Each of these constructs contained Osem promoter fragments of different lengths and were fused to the 5′ end of the CaMV 35S minimal promoter and connected to the β-glucuronidase (GUS) reporter gene. The 101-bp-TATA and 55-bp-TATA constructs were, respectively, driven by Osem promoter fragments from –176 to –76, and from –179 to –125 (Fig. 1b). These were tested for ABA induction and VP1 activation in a rice protoplast transient assay system. As shown in Fig. 1(c,d), high levels (67-fold for the 101-bp-TATA and 55-fold for the 55-bp-TATA construct) of ABA induction were observed, although the absolute level of GUS activity obtained with the 101-bp-TATA was significantly greater than that obtained with the 55-bp-TATA construct. Similarly, remarkable levels of VP1 activation were observed for both constructs. Although the constructs used in the experiments described in Fig. 1 and elsewhere in this work contained an intron derived from a castor bean catalase gene, we observed similar levels of ABA induction and VP1 activation with the constructs in which the intron had been removed (data not shown). This contrasts with results reported for the ABRC in the barley HVA22 promoter, which requires the presence of an intron for ABA induction (Shen & Ho 1995). Since the 55 bp fragment (which contained motif A, region 1 and the CE3 sequence) was reasonably short and was found to be sufficient to confer both ABA and VP1 responsiveness to the minimal promoter, we further analyzed this fragment by linker-scan and base-substitution mutagenesis, as described below.

ABA responsiveness in the 55 bp fragment requires a cis-acting motif A and CE3 sequence

A series of linker-scan mutants of the 55-bp-TATA construct with sequence replacements at 10 bp intervals were generated and tested for ABA induction. As shown in Fig. 2(a), sequence replacements in motif A (mut-1) and of the CE sequence (mut-4) strikingly reduced the level of ABA induction. Some decrease in the extent of induction was also observed for mut-2, in which the 10 bp sequence just downstream of motif A was replaced. In contrast, the third (mut-3) and fifth (mut-5) mutations enhanced rather than reduced the level of induction. The substitution of four nucleotides in the 5 bp sequence at the 5′ end of the 55 bp (mut-0) did not reduce the extent of ABA induction (Fig. 2b). These results indicate that the ACGT-containing element (motif A) and the CE3 sequence together here function in the same manner as the ABRE (A2) and coupling element (CE3) of ABRC3 in the barley HVA1 promoter.

image

Figure 2. Linker-scan analysis for ABA induction of the 55 bp fragment of the Osem promoter.

The 55-bp-TATA (wt) construct (schematically illustrated at the top) and the linker-scan derivatives [mut-1 to –5 (a); mut-0 (b)] were tested for ABA-induction as in Fig. 1.

The nucleotide sequences of the 55 bp regions of the constructs are shown on the left. Dashes (–) mark nucleotides that match wild-type sequence, and lower case letters indicate those that differ. Values of GUS activities normalized to LUC activities are relative to those obtained with the wild-type 55-bp-TATA without ABA and are the means of three independent electroporations with standard errors. Results shown in (a) and (b) are from different experiments with independent protoplast preparations.

Download figure to PowerPoint

The effect of the mut-2 mutation may indicate the presence of another cis-acting element that quantitatively affects the level of ABA induction. Alternatively, the recognition sequence of the factor that binds to motif A may extend to the mut-2 mutation site.

VP1 responsiveness in the 55 bp fragment also requires both motif A and the CE3 sequence

The cis-acting elements required for VP1 activation mediated by the 55 bp promoter fragment were defined by testing VP1 responsiveness of the linker-scan mutations described above (Fig. 3). Among the five mutants of the linker-scan series, mut-1, in which motif A was replaced, almost completely lost its capacity to be activated by VP1. The CE3 mutant, mut-4, also exhibited a marked reduction in VP1 activation. However, the effect of the mutation was not as profound as that of the mut-1 mutation. Surprisingly, decreases in the level of VP1 activation were observed with the other three mutants as well, although the extent of decrease was somewhat smaller than that observed with mut-4. A decrease in the extent of activation with mut-5, however, instead appeared due to an increase in the level of basal expression. Mutation of the sequence upstream of motif A (mut-0) had a little effect on activation, but the absolute level of expression was higher than that of the wild-type (Fig. 3b). These results demonstrate that not only motif A but also the CE3 sequence is important for the 55 bp fragment to be activated by VP1. Motif A appears to be the more critical of the two, however. Since the mut-2 and mut-3 mutants exhibited decreases in the level of VP1 activation, the region spanning these mutation sites probably contains other cis-acting elements that operate in combination with motif A and CE3. This was confirmed by the further loss of VP1 activation (Fig. 3b), which occurred when the 30 bp region downstream of motif A was fully replaced with an unrelated A/T-rich sequence. The effect of the mut-2 mutation on VP1 activation may be related to the decrease in the level of ABA induction with the same mutant.

image

Figure 3. Linker-scan analysis of the 55 bp fragment of the Osem promoter for VP1 activation.

The 55-bp-TATA (wt) construct (schematically illustrated at the top) and the linker-scan derivatives [(a), mut-1 to –5; (b), mut-0 and –2/3/4)] were tested for VP1 activation as in Fig. 1. Other aspects of experiments and analyses were as described for Fig. 2.

Download figure to PowerPoint

The ACGT-containing ABRE and the CE3 coupling element have a qualitatively equivalent function in ABA-induced transcription

The possibility that motif A and CE3 could be interchangeable in the 55 bp context arose from our observation that a 6 bp portion of motif A was identical to the CE3 sequence (CGTGTC). This hypothesis was tested by replacing each sequence with the other, as shown in Fig. 4, such that motif A was replaced with CE3 (mut-BB), CE3 with motif A (mut-AA), or both being replaced by each other (mut-BA). All three exhibited ABA induction comparable to that of the wild-type 55-bp-TATA construct (Fig. 4a). These results indicate that motif A and CE3 are functionally equivalent with respect to the ABA-responsiveness of the 55 bp cis-acting element complex.

image

Figure 4. Motif-change experiments testing functional equivalence of motif A and the CE3 sequence.

Motif-exchange mutants of the 55-bp-TATA construct indicated on the left were tested for ABA induction (a) and VP1 activation (b). Results shown in (a) and (b) are from different experiments with independent protoplast preparations. Other aspects of experiments and analyses were as described for Figs 2 and 3.

Download figure to PowerPoint

The same motif-exchange mutants were tested for VP1 activation. As shown in Fig. 4(b), all three mutants were clearly activated by VP1 although a decrease to some degree in the level of activation was observed when the CE3 sequence was replaced with motif A (mut-AA) or when they were exchanged (mut-BA). Therefore, motif A and the CE3 sequence are functionally similar in VP1 activation as well as ABA-responsiveness, although for VP1 activation the efficiency with motif A placed at the CE3 position was lower than that of the original CE3. The results of these motif-exchange experiments strongly suggest that the similarity in the sequence between motif A and CE3 reflects the functional equivalence both in ABA induction and in VP1 activation.

The CE3 element can also be specifically recognized by TRAB1, a VP1-interacting ABRE-binding factor

Since the CE3 element and ACGT-containing ABRE (motif A) were similar in sequence and functionally interchangeable, we hypothesized that the CE3 sequence can bind the same factor that acts on ACGT-containing ABREs. To test this hypothesis, we conducted an electrophoretic mobility shift assay using in vitro translated TRAB1, an ABRE-binding bZIP factor that interacts with VP1 and can mediate an ABA signal. As shown in Fig. 5, TRAB1 complexed not only with the wild-type 55 bp fragment but also with the motif A (mut-1) or CE3 (mut-4) mutant. In contrast, no specific protein–DNA complex was observed with the fragment in which both motif A and CE3 were replaced (mut-1/4). In addition, TRAB1 bound to an isolated CE3 sequence independently of motif A (Fig. 6b). Futhermore, the TRAB1 binding to motif A was efficiently competed out by the CE3 sequence (Fig. 6c). These results indicate that the CE3 sequence and motif A are specifically bound by the same trans-acting factor.

image

Figure 5. Electrophoretic mobility shift assay testing whether TRAB1 binds to the CE3 sequence.

A protein synthesis reaction in a rabbit reticulocyte lysate programmed with TRAB1 mRNA (+ TRAB1; lane 1–5) or without added RNA templates (– TRAB1; lane 6–10) was incubated with 32P-labeled double-stranded oligonucleotides of indicated sequences (lanes 1 and 6, wt; lanes 2 and 7, mut-1; lanes 3 and 8, mut-3/4; lanes 4 and 9, mut-4; lanes 5 and 10, mut-1/4). DNA–protein complexes were separated from free DNAs by non-denaturating polyacrylamide gel electrophoresis. Positions of protein–DNA complexes specific to TRAB1 are indicated by an arrow. Other bands derive from the protein present in the reticulocyte lysate.

Download figure to PowerPoint

image

Figure 6. Electrophoretic mobility shift assay testing whether TRAB1 binds to non-ACGT ABREs.

(a) The nucleotide sequences of the probes used.

(b) Electrophoretic mobility shift assays were conducted as in Fig. 5 using ACGT-containing or non-ACGT ABRE sequences as indicated.

(c) Competition assays using the motif A sequence as probe. Binding reactions were performed with an indicated unlabeled competitor. ‘x’ indicates fold-molar excess of the added competitor. Other details are as described for Fig. 5.

Download figure to PowerPoint

TRAB1 binds to other non-ACGT ABREs

The results described above indicate that CE3 itself has capacity to mediate ABA signals. Thus, CE3 can be referred to as a non-ACGT ABRE. Two other non-ACGT ABREs, which can also intrinsically mediate ABA signals, are known. One is motif III (GCCGCGTGGC), which has been identified in the Rab16B promoter and can confer ABA-responsiveness to the CaMV 35S minimal promoter when tetramerized (Ono et al. 1996). The other is hex3 (GACGCGTGGC), which was artificially created as a mutant sequence of the hexamer motif found in the histone promoters. A tetramer of hex3 has also been demonstrated to confer ABA-responsiveness to the CaMV 35S–90 promoter (Lam & Chua 1991). Although both sequences, along with CE3, have been recognized as an element different from ACGT-containing ABREs (Lam & Chua 1991; Ono et al. 1996), we noticed a common feature among these sequences. The pentanucleotide GTGGC or GTGTC is known as a strong half-site of the palindromic G-box sequence that binds GBF-type or class I bZIP proteins (Izawa et al. 1993; Niu & Guiltinan 1994). As illustrated in Fig. 7, the CE3 sequence (A/G)-4A–3C–2G–1C0G0T+1G+2T+3C+4 is regarded as a near-complete palindrome of one of the strong half-sites that differs at only the –1 position (Fig. 7). Similarly, motif III is a near complete palidrome of another strong half site also with an A to G substitution at –1. Hex3 is regarded as a hybrid of the two strong half sites with the same base substitution. These observations led us to test whether TRAB1 can specifically bind to these non-ACGT ABREs.

image

Figure 7. Sequence similarity among ACGT-containing and non-ACGT ABREs including CE3 sequences.

(a) The palindromes of the two strong half sites of G-box sequences for GBF-type bZIP protein binding are shown.

(b) Sequences of ACGT-containing and non-ACGT ABREs are compared. Arrows on the top and bottom of each sequence indicate the identities with each strong half-site shown in (a). Two types of strong half-sites shown in (a) are differentiated by solid and broken lines. Nucleotides different from those in the sequences of strong half-sites are indicated by lowercase letters. The ACGT-cores or the corresponding nucleotides are italicized.

Download figure to PowerPoint

As shown in Fig. 6(b,c), TRAB1 was also found to bind to motif III and hex3 with an equal or even higher affinity as compared to motif A. These results indicate that these non-ACGT ABREs, as well as CE3, are the same type of cis-element as ACGT-containing ABREs.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The ACGT-containing ABRE and CE3 are required for both ABA and VP1 regulation

In the present study, we first defined a short fragment of the Osem promoter that can confer high levels of ABA induction and VP1 activation to a heterologous minimal promoter, then analyzed this construct to define the elements required for both ABA and VP1 responsiveness. These analyses have demonstrated that, as in the HVA1 promoter, the CE3 sequence present in the 55 bp fragment functions as a coupling element for ABA-responsiveness. This CE3 sequence also acts together with the ACGT-containing ABRE (motif A) located 30 bp upstream. Although the CE3 element is located just upstream of the ACGT-containing ABRE in the ABRC3, in the HVA1 promoter, Shen et al. (1996) have shown that it can be artificially placed downstream of the ABRE in a similar distance to that which separates motif A and the CE3 sequence in the Osem promoter.

Contrary to the present results, we previously reported that a 31 bp internal deletion, which removed the CE3 sequence, in the 502 bp promoter context had little effect on ABA-inducibility. One possible explanation for this apparent discrepancy is that the deletion might have created a new ABRC. For example, the downstream ACGT-containing ABRE (motif A′) or another potentially as yet unidentified coupling element was brought to the vicinity of motif A by the deletion and might have become coupled to motif A. Other cryptic coupling elements might also exist.

We have previously shown that motif A plays a critical role not only in ABA induction but also in VP1 activation (Hattori et al. 1995). The importance of ACGT-containing ABREs in VP1 activation has also been demonstrated by Vasil et al. (1995). These authors have further shown that the presence of multiple G-box sequences is sufficient for VP1 activation (Vasil et al. 1995). They were able to confer VP1 activation to a heterologous minimal promoter by using a tetramer of 22 bp sequences each containing a G-box with its flanking regions from the parsley CHS promoter. However, it was not clear whether or not a single copy of a G-box or ACGT-containing sequence is sufficient to mediate VP1 activation as well as ABA-inducibility in a natural promoter context. In addition, it has not previously been established whether the coupling element plays any role in VP1 activation. The present study has clearly demonstrated that the presence of a single copy of a ACGT-containing ABRE alone is not sufficient for VP1 activation. The strongest supporting evidence was apparent in the results from mut-2/3/4 (Fig. 3b). In addition, the work here has shown that the CE3 sequence and other element(s) lying between motif A and CE3 are also necessary for VP1 responsiveness in the Osem promoter.

Sequence and functional equivalence between the ACGT-containing ABRE and the CE3 coupling element

In the present study, we have demonstrated that the ACGT-containing ABRE and the CE3 coupling element are functionally interchangeable. Such functional equivalence is considered to be based on their common sequence structure (CGTG(G/T)C) and binding of the same trans-acting factor (TRAB1).

In addition to CE3, we have demonstrated that the ABA-responsive TRAB1 can also bind to two other non-ACGT ABREs, motif III and hex3, which have previously been taken as ABREs distinct from the ACGT-containing type. Based on our results here, we propose that these non-ACGT ABREs, including CE3, and ACGT containing ABREs should be categorized into a single class of cis-elements. For convenience, we use here ‘G-box-related ABRE’ (G-ABRE) for this class of the ABA-responsive sequence, since a completely different class of cis-elements that can intrinsically mediate ABA signals might be found in the future. The sequence features of G-ABRE are: (i) it contains at least one complete strong half site of GBF binding (Izawa et al. 1993; Niu & Guiltinan 1994; see Fig. 7); (ii) the ACGT-type features the ACGT core by extending AC to 5′ of the strong half site; and (iii) non-ACGT-type is either a palindoreme or a hybrid of the strong half sites with a single base change at –1 position, which corresponds to ‘A’ of the ACGT core (Fig. 7). Although the ACGT sequence is a typical core and an important sequence for the binding of plant bZIP proteins, the possible negative effect of a single-base deviation in this sequence may be compensated, in the present instance, by the nearly complete match to the palindrome of the strong half-sites. Other work has also shown the binding of GBF-type bZIP proteins to near-plindromic, non-ACGT sequences (Hong et al. 1995; Niu & Guiltinan 1994; Schindler et al. 1992).

Coupling elements

The coupling element has been proposed to specify ABA-responsiveness in combination with an ACGT-containing ABRE (Shen & Ho 1995; Shen et al. 1996). However, our results indicate that the CE3 coupling element functions instead as a second G-ABRE, which is recognized by an identical trans-acting factor. Therefore, two copies of G-ABREs, either the ACGT-type or the non-ACGT (CE3)-type, which are arranged within an appropriately short distance, appear to constitute an ABRC in the natural promoter context. This can explain the observation that a single copy of ABRE is not sufficient for ABA induction and that only artificially multimerized ABREs can confer ABA-responsiveness to a heterologous promoter. In the wheat Em promoter, both of the two ACGT-containing ACGT-containing ABREs, namely Em1a and Em1b, which are separated by 45 bp, are essential for ABA induction in the –243 promoter context (Vasil et al. 1995). Thus, Em1a and Em1b (Marcotte et al. 1989) may constitute an ABRC (complex 1; Guiltinan et al. 1990) which does not contain the CE3-type sequence.

In contrast to CE3, the sequence of the coupling element CE1 appears to be markedly different from the ACGT-containing ABRE (Shen & Ho 1995). Therefore, a single copy of the ACGT-containing ABRE can also form an ABRC with a distinctly different type of element. These two types of ABRCs (two copies of G-ABREs or G-ABRE + CE1) appear to be qualitatively different with respect to VP1 responsiveness. The cis-acting ABRCs composed of two or more G-ABRE sequences all appear to be activated by VP1. These include the ABRC3 of HVA1 (Shen et al. 1996), the 55 bp fragment and its motif-exchanged derivatives of Osem (this study), complex 1 of the wheat Em (Guiltinan et al. 1990), the Rab28 promoter (Busk & Pages 1997) and the artificially multimerized G-box sequences (Vasil et al. 1995). In contrast, ABRC1 of HVA22 composed of a single copy of a ACGT-containing ABRE and CE1 has been demonstrated not to be activated by VP1 (Shen et al. 1996). From these observations, we propose that two copies of G-ABREs sequences constitute a VP1-regulated ABA-responsive cis-acting element complex.

From our and other results, it is evident that G-ABREs intrinsically can mediate the ABA signal and do not necessarily require a coupling element of distinct class even in a natural promoter context. In other words, G-ABREs can ‘self-couple’. Thus, the two copies of the G-ABREs appear to function in a highly synergistic manner in the transcriptional activation by ABA. Studies in mammalian systems have shown that synergistic actions of multiple activators are required for the assembly or recruitment of general transcription factors (Chi et al. 1995). Therefore, the synergy between the G-ABREs examined here may be achieved through the action on the basic transcription machinery. A similar synergism may be also be involved in two-element combinations of a G-ABRE and dissimilar sequence such as CE1. Since the effect of the two G-ABREs is so highly synergistic, co-operative interactions may also exist between the trans-acting factors or complexes that bind to the G-ABRE elements.

Complexity of the cis-acting 55 bp fragment of the Osem promoter

VP1 functions through G-ABREs by interacting with an ABRE-binding bZIP protein such as TRAB1. In addition, the presence of two copies of G-ABREs is an essential feature of the 55 bp, VP1-regualted ABRC of Osem. However, it would be an oversimplification to conclude that they are sufficient to account for the full function of the 55 bp ABRC. The mutation just downstream of motif A (mut-2) quantitatively affected the level of ABA induction and VP1 activation. The latter was also observed in the mut-3 mutation. These results may be related to the previous observation that internal deletion of a 17 bp sequence between motif A and the CE3 sequence in the full promoter context resulted in a great reduction of the level of ABA induction and VP1 activation (Hattori et al. 1995). However, the possibility cannot be excluded that the effect of this internal deletion could have been due, instead, to a change in spacing between motif A and the CE3 sequence.

Considering the mechanism that VP1 acts through G–ABREs via interaction with G-ABRE-binding factors, VP1 regulation is considered to be tightly coupled with the ABA signal. In addition, in maize embryo, VP1 is an absolute requirement for ABA-induced expression of Em (homologue) (McCarty et al. 1991). Thus, one might expect that promoter mutations which reduce the level of VP1 activation would affect the ABA-induced transcription. However, the results from mut-3 showed that this was not the case; the mutation did not reduce the level of ABA induction despite its decreasing effect on the level of VP1 activation. This may suggest the presence of an additional element through which VP1 acts independently of ABA-regulation. We also reported previously that mutation of an Sph box in the Osem promoter reduced VP1 activation but affected ABA-induction very little. In other respects, the VP1-independent ABA-induction mechanism may operate at the same time for the 55 bp ABRC in our rice cell system, as is the case for the induction of ABRC3 in vegetative tissues which do not express VP1 (Shen et al. 1996). Similarly, the maize Rab28 can be induced by ABA in excised VP1-deficient mutant embryo, although its expression during normal embryo development was significantly low in the vp1 mutant (Pla et al. 1991). However, the results of transient co-transfection assays should be interpreted with caution. The ABA-response of the rice cell-line used for the protoplast transient assay could be similar to those of embryos because OSVP1 is expressed at an equal or higher level (Nakagawa et al. 1996a). However, it also means that the co-transfection experiments with 35S-Sh-Osvp1 in the present study only measured the effect of OSVP1 overexpression. This may pose a limitation on observing a tight coupling between ABA and VP1 regulation using our protoplast system.

Although it is clear that motif A and the CE3 sequence intrinsically have equivalent function, relative contribution by the two elements to the VP1 activation appeared to differ in the 55 bp context: the effect of a mutation in motif A was more profound than in CE3. This suggests that factors or complexes of factors acting together with VP1 that bind to ABRE/CE3 elements may differentially interact with other elements and/or the basic transcription machinery.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Transient expression in protoplasts

Transient expression experiments by electroporation using protoplasts from a rice cultured cell-line (Oc) were carried out essentially as described by Hattori et al. (1995). For ABA-induction experiments, protoplast samples (4 × 106 cells) were electroporated with 10 μg of the test plasmid carrying the GUS reporter gene and 10 μg of a ubiquitin promoter-luciferase plasmid as an internal standard (Christensen et al. 1992). After electroporation, each sample was divided into two equal portions, each of which was cultured for 40 h in 3 ml of R2P medium (Kyozuka & Shimamoto 1991) either with or without 5 × 10–5 M ABA. For VP1-activation experiments, protoplast samples were electroporated with 20 μg of the 35S-Sh-Osvp1 effector plasmid (Hattori et al. 1994) in addition to the GUS and luciferase plasmids. The 35S-Sh-none plasmid, in which the Osvp1 sequence was deleted from that of 35S-Sh-Osvp1, was used as the control for the effector plasmid. Each treatment was repeated three times in experiments using single preparations of protoplasts. After incubation, protoplasts were extracted with buffer as in Lanahan et al. (1992) and used for GUS assays as described previously (Hattori et al. 1995) as well as luciferase assays using the Pikkagene luciferase assay kit (Promega, Madison, WI, USA). The GUS activities reported were normalized to the luciferase activity.

Plasmid construction

pIG46 (Ono et al. 1996), which consists of the CaMV 35S minimal promoter, the castor bean CAT1 first intron and the GUS gene, was used to make various chimeric promoter-GUS constructs. The HindIII and XhoI sites of pIG46 located at the 5′-end of the minimal promoter were used to inserted various promoter fragments. pIG46 was first digested with XhoI, blunt-ended by the Klenow fragment of DNA polymerase I and then digested with HindIII. This linearized plasmid was ligated with an HindIII/NaeI fragment of the Osem promoter which spans from –176 to –76, and was prepared from a d2 plasmid (Hattori et al. 1995) to produce the 101-bp-TATA constructs. The 55-bp-TATA construct and its linker-scan and other mutant derivatives were prepared by ligating HindIII/XhoI-digested pIG46 and double-stranded synthetic oligonucleotides having the sequences shown in Figs 2, 3 and 4 and 5′-overhangs compatible with HindIII and XhoI cut ends. Parts of the constructs, which derive from synthetic DNAs, were all sequenced.

Electrophoretic mobility shift assays

An EcoRI-SalI fragment of TRAB1 cDNA (the entire fragment) was cloned into EcoRI/SalI-digested pCITE-4a (Novagen, Madison, WI, USA), linearized with SpeI (3′ to the SalI site), and transcribed with T7 RNA polymerase. The transcribed RNA was translated in a rabbit reticulocyte lysate (Promega, Madison, WI, USA) according to the manufacturer’s instructions. A 3 μl aliquot of the translation reaction was included in a 20 μl DNA-binding reaction and resolved on a 5% polyacrylamide gel as described previously (Nakagawa et al. 1996b). Control binding reactions were performed using translation reaction media without template. The nucleotide sequences of the double-stranded synthetic oligonucleotides used as probes were indicated in each figure.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We wish to thank Dr Karen E. Koch (University of Florida, Gainesville, FL, USA) for critical reading of the manuscript. This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (‘Molecular Basis of Flexible Organ Plans in Plants’, no. 06278102) from the Ministry of Education, Science and Culture, Japan, and by a project grant from the Ministry of Agriculture, Forestries and Fisheries, Japan.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Block, A., Dangl, J.L., Hahlbrock, K. & Schulze-Lefert, P. 1990 Functional borders, genetic fine structure, and distance requirements of cis elements mediating light responsiveness of the parsley chalcone synthase promoter. Proc. Natl Acad. Sci. USA, 87, 53875391.
  • Bobb, A.J., Eiben, H.G. & Bustos, M.M. 1995 PvAlf, an embryo-specific acidic transcriptional activator enhances gene expression from phaseolin and phytohemagglutinin promoters. Plant J., 8, 331343.
  • Busk, P.K., Jensen, A.B. & Pages, M. 1997 Regulatory elements in vivo in the promoter of the abscisic acid responsive gene rab 17 from maize. Plant J., 11, 12851295.
  • Busk, P.K. & Pages, M. 1997 Protein binding to the abscisic acid-responsive elements is independent of VIVIPROUS1 in vivo. Plant Cell, 9, 22612270.
  • Busk, P.K. & Pages, M. 1998 Regulation of abscisic acid-induced transcription. Plant Mol. Biol., 37, 425435.
  • Carson, C.B., Hattori, T., Rosenkrans, L., Vasil, V., Vasil, I.K., Peterson P.P. & McCarty, D.R. 1997 The quiescent/colorless allele of viviparous1 show that the conserved B3 domain of VP1 is not essential for ABA-regulated gene expression in the seed. Plant J., 12, 12311240.
  • Chi, T., Lieberman, P., Ellwood, K. & Carey, M. 1995 A general mechanism for transcriptional synergy by eukaryotic activators. Nature, 377, 254257.
  • Christensen, A.H., Sharrock, R.A. & Quail, P.H. 1992 Maize polyubiquitin genes: structure, thermal perturbation of expression and transcript splicing, and promoter activity following transfer to protoplasts by electroporation. Plant Mol. Biol., 18, 675689.
  • Giraudat, J., Hauge, B.M., Valon, C., Smalle, J., Parcy, F. & Goodman, H.M. 1992 Isolation of the Arabidopsis ABI3 gene by positional cloning. Plant Cell, 4, 12511261.
  • Giraudat, J., Parcy, F., Bertauche, N., Gosti, F., Leung, J., Morris, P.C., Bouvier-Durand, M. & Vartanian, N. 1994 Current advances in abscisic acid action and signalling. Plant Mol. Biol., 26, 15571577.
  • Giuliano, G., Pichersky, E., Malik, V.S., Timko, M.P., Scolnik, P.A. & Cashmore, A.R. 1988 An evolutionarily conserved protein binding sequence upstream of a plant light-regulated gene. Proc. Natl Acad. Sci. USA, 85, 70897093.
  • Guiltinan, M.J., Marcotte, W.R. & Quatrano, R.S. 1990 A plant leucine zipper protein that recognizes an abscisic acid response element. Science, 250, 267270.
  • Hattori, T., Terada, T. & Hamasuna, S. 1994 Sequence and functional analyses of the rice gene homologous to the maize. Vp1. Plant Mol. Biol., 24, 805810.
  • Hattori, T., Terada, T. & Hamasuna, S. 1995 Regulation of the Osem gene by abscisic acid and the transcriptional activator VP1: analysis of cis-acting promoter elements required for regulation by abscisic acid and VP1. Plant J., 7, 913925.
  • Hattori, T., Vasil, V., Rosenkrans, L., Hannah, L.C., McCarty, D.R. & Vasil, I.K. 1992 The Viviparous-1 gene and abscisic acid activate the C1 regulatory gene for anthocyanin biosynthesis during seed maturation in maize. Genes Dev., 6, 609618.
  • Hong, J.C., Cheong, Y.H., Nagao, R.T., Bahk, J.D., Key, J.L. & Cho, M.J. 1995 Isolation of two soybean G-box binding factors which interact with a G-box sequence of an auxin-responsive gene. Plant J., 8, 99211.
  • Izawa, T., Foster, R. & Chua, N.-H. 1993 Plant b-ZIP protein DNA-binding specificity. J. Mol. Biol., 230, 11311144.
  • Kao, C.Y., Cocciolone, S.M., Vasil, I.K. & McCarty, D.R. 1996 Localization and interaction of the cis-acting elements for abscisic acid, VIVIPAROUS1, and light activation of the C1 gene of maize. Plant Cell, 8, 11711179.
  • Kyozuka, J. & Shimamoto, K. 1991 Transformation and regeneration of rice protoplasts. In Plant Tissue Culture Manual, B1 (Lindsey, K., ed.) Dordrecht. Kluwer Academic Publishers, pp. 116.
  • Lam, E. & Chua, N.-H. 1991 Tetramer of a 21-base pair synthetic element confers seed expression and transcriptional enhancement in response to water stress and abscisic acid. J. Biol. Chem., 266, 1713117135.
  • Lanahan, M.B., Ho, T.-H., Rogers, S.W. & Rogers, J.C. 1992 A gibberellin response complex in cereal alpha-amylase gene promoters. Plant Cell, 4, 203211.
  • Liu, Z.B., Ulmasov, T., Shi, X., Hagen, G. & Guilfoyle, T.J. 1994 Soybean GH3 promoter contains multiple auxin-inducible elements. Plant Cell, 6, 645657.
  • Marcotte, W.R. Jr, Russel, S.H. & Quatrano, R.S. 1989 Abscisic acid-responsive sequences from the Em gene of wheat. Plant Cell, 1, 969976.
  • Mason, H.S., DeWald, D.B. & Mullet, J.E. 1993 Identification of a methyl jasmonate-responsive domain in the soybean vspB promoter. Plant Cell, 5, 241251.
  • McCarty, D.R., Hattori, T., Carson, C.B., Vasil, V. & Vasil, I.K. 1991 The Viviparous-1 developmental gene of maize encodes a novel transcriptional activator. Cell, 66, 895905.
  • Meshi, T. & Iwabuchi, M. 1995 Plant transcription factors. Plant Cell Physiol., 36, 4051420.
  • Michel, E., Salamini, R., Barels, E., Dale, P., Baga, M. & Szalay, A. 1993 Analysis of a desiccation and ABA-responsive promoter isolated from the resurrection plant Craterostigma plantagineum. Plant J., 4, 2940.
  • Mundy, J., Yamaguchi-Shinozaki, K. & Chua, N.-H. 1990 Nuclear proteins bind conserved elements in the abscisic acid-responsive promoter of a rice rab gene. Proc. Natl Acad. Sci. USA, 87, 14061410.
  • Nakagawa, H., Ohkura, E., Ohmiya, K. & Hattori, T. 1996a Seed specific transcription factor VP1 (OSVP1) is expressed in rice suspension-cultured cells. Plant Cell Physiol., 37, 355362.
  • Nakagawa, H., Ohmiya, K. & Hattori, T. 1996b A rice bZIP protein, designated OSBZ8, is rapidly induced by abscisic acid. Plant J., 9, 217227.
  • Nakayama, T., Ohtsubo, N., Mikami, K., Kawata, T., Tabata, T., Kanazawa, H. & Iwabuchi, M. 1989 Cis-acting sequences that modulate transcription of wheat histone H3 gene and 3′ processing of H3 premature mRNA. Plant Cell Physiol., 30, 825832.
  • Niu, X. & Guiltinan, M.J. 1994 DNA binding specificity of the wheat bZIP protein EmBP-1. Nucl. Acids Res., 22, 49694978.
  • Oeda, K., Salinas, J. & Chua, N.-H. 1991 A tobacco bZip transcription activator (TAF-1) binds to G-Box-like motif conserved in plant genes. EMBO J., 10, 17931802.
  • Ono, A., Izawa, T., Chua, N.-H. & Shimamoto, K. 1996 The rab16B promoter of rice contains two distinct abscisic acid-responsive elements. Plant Physiol., 112, 483491.
  • Parcy, T., Valon, C., Raynal, M., Gaubier-Comell, P., Delseny, M. & Giraudat, J. 1994 Regulation of gene expression programs during Arabidopsis seed development: Roles of the ABI3 locus and of endogenous abscisic acid. Plant Cell, 6, 15671582.
  • Pla, M., Vilardell, J., Guiltinan, M.J., Marcotte, W.R., Niogret, M.F., Quatrano, R.S. & Pages, M. 1993 The cis-regulatory element CCACGTGG is involved in ABA and water-stress responses of the maize gene. Rab28. Plant Mol. Biol., 21, 259266.
  • Schindler, U., Terzaghi, W., Beckman, H., Kadesch, T. & Cashmore, A.R. 1992 DNA-binding site preferences and transcriptional activation properties of the Arabidopsis transcription factor GBF1. EMBO, J., 11, 12751289.
  • Schulze-Lefert, P., Dangl, J.L., Becker-Andre, M., Hahlbrock, K. & Schulz, W. 1989 Inducible in vivo DNA footprints define sequences necessary for UV light activation of the parsley chalcone synthase gene. EMBO J., 8, 651656.
  • Shen, Q. & Ho, T.-H.D. 1995 Functional dissection of an abscisic acid (ABA) -inducible gene reveals two independent ABA-responsive complexes each containing a G-box and novel cis-acting element. Plant Cell, 7, 295307.
  • Shen, Q., Zhang, P. & Ho, T.-H.D. 1996 Modular nature of abscisic acid (ABA) response complexes: composite promoter units that are necessary and sufficient for ABA induction of gene expression in barley. Plant Cell, 8, 11071119.
  • Skriver, K., Olsen, F.L., Roger, J. & Mundy, J. 1991 Cis-acting DNA elements responsive to gibberellin and its antagonist abscisic acid. Proc. Natl Acad. Sci. USA, 88, 72667270.
  • Suzuki, M., Kao, C.Y. & McCarty, D.R. 1997 The conserved B3 domain of VIVIPAROUS1 has a cooperative DNA binding activity. Plant Cell, 9, 799807.
  • Ulmasov, T., Liu, Z.B., Hagen, G. & Guilfoyle, T.J. 1995 Composite structure of auxin response elements. Plant Cell, 7, 16111623.
  • Vasil, V., Marcotte, W.R. Jr, Rosenkrans, L., Cocciolone, S.M., Vasil, I.K., Quatrano, R.S. & McCarty, D.R. 1995 Overlap of Viviparous1 (VP1) and abscisic acid response elements in the Em promoter: G-box elements are sufficient but not necessary for VP1 transactivation. Plant Cell, 7, 15111518.
  • Williams, M.E., Foster, R. & Chua, N.-H. 1992 Sequences flanking the hexameric G-box core CACGTG affect the specificity of protein binding. Plant Cell, 4, 485496.
  • Yamaguchi-Shinozaki, K., Mundy, J. & Chua, N.-H. 1989 Four tightly linked rab genes are differentially expressed in rice. Plant Mol. Biol., 14, 2939.