Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
The redundancy of genes for plant transcription factors often interferes with efforts to identify the biologic functions of such factors. We show here that four different transcription factors fused to the EAR motif, a repression domain of only 12 amino acids, act as dominant repressors in transgenic Arabidopsis and suppress the expression of specific target genes, even in the presence of the redundant transcription factors, with resultant dominant loss-of-function phenotypes. Chimeric EIN3, CUC1, PAP1, and AtMYB23 repressors that included the EAR motif dominantly suppressed the expression of their target genes and caused insensitivity to ethylene, cup-shaped cotyledons, reduction in the accumulation of anthocyanin, and absence of trichomes, respectively. This chimeric repressor silencing technology (CRES-T), exploiting the EAR-motif repression domain, is simple and effective and can overcome genetic redundancy. Thus, it should be useful not only for the rapid analysis of the functions of redundant plant transcription factors but also for the manipulation of plant traits via the suppression of gene expression that is regulated by specific transcription factors.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The genome of Arabidopsis contains more than twice as many genes for transcription factors as the genomes of animals with genomes of sizes similar to that of Arabidopsis (Riechmann et al., 2000). In addition, the transcription factors encoded in plant genomes exhibit much greater variety than those in animals, and numerous plant-specific transcription factors have been recognized (Riechmann et al., 2000). However, the functions and target genes of most plant transcription factors remain to be characterized. Plant genes are frequently duplicated, and many plant transcription factors form large families in which family members include strongly conserved DNA-binding domains (Riechmann et al., 2000). This structural and functional redundancy of plant transcription factors often interferes with efforts to identify the functions of these factors. Even when gene-knockout or antisense lines specific for a particular transcription factor can be isolated, such lines often fail to exhibit an informative phenotype that might provide some direct clue to the factor's function, for example, a loss-of-function phenotype (Borevitz et al., 2000; Bouché and Bouchez, 2001; Meissner et al., 1999).
In animal cells, chimeric repressors, in which a DNA-binding domain or a transcription factor is fused to a repression domain, have been used for the targeted repression of the expression of genes of interest (Badiani et al., 1994; Beerli et al., 1998, 2000; de Haan et al., 2000; John et al., 1995). The repression domains from the Krüppel-associated box (KRAB) of the human estrogen receptor, Engrailed (En) of Drosophila or mSIN3 interaction domain (Sid) of humans, have been used effectively in mammalian cells (Badiani et al., 1994; Beerli et al., 1998, 2000; de Haan et al., 2000; John et al., 1995). Such chimeric repressors have been shown to act dominantly in animals, and thus a similar strategy might be expected to facilitate the analysis of redundant transcription factors in plants. However, the repressive activity of animal repression domains has not been demonstrated clearly in plant cells. In fact, the KRAB repression domain acts as a transcriptional activator in transient expression assays in Arabidopsis when fused to the GAL4 DNA-binding domain (unpublished results). This observation suggests that the mechanisms for the repression of transcription might differ between animals and plants. Moreover, addition of a long peptide, such as En, which includes approximately 300 amino acids, might affect the integral properties of some plant transcription factors, such as protein–protein interactions and/or dimerization. Thus, repression domains that function in plant cells and can overcome the genetic redundancy of plant transcription factors are necessary if we are to utilize chimeric repressors or the identification of the biologic functions of plant transcription factors.
We previously reported that the repression domains of the class II ETHYLENE-RESPONSIVE ELEMENT-BINDING FACTOR (ERF) and TFIIIA-type zinc finger repressors of transcription that include SUPERMAN (SUP) contain the EAR (ERF-associated amphiphilic repression) motif (Hiratsu et al., 2002; Ohta et al., 2001). When short peptides that contained the EAR motif were fused to activators of transcription, the resultant chimeric transcription factors acted as strong repressors and suppressed the expression of a reporter gene in the presence of another activator of transcription in transient expression assays in Arabidopsis, even if the activator contained the VP16 activation domain of herpes simplex virus (Hiratsu et al., 2002; Ohta et al., 2001). Given that the EAR motif was able to convert a transcriptional activator into a strong repressor, and that the repressive activity of the EAR-motif repression domain was dominant over both intra- and intermolecular activational activities (Hiratsu et al., 2002; Ohta et al., 2001), we attempted, in the present study, to convert intact plant transcription factors into dominant repressors by fusing them with the EAR-motif repression domain and to suppress the expression of the target genes of the various transcription factors in stably transformed Arabidopsis.
We chose four transcription factors from three different families, namely the products of the ETHYLENE-INSENSITIVE3 (EIN3), CUP-SHAPED COTYLEDON1 (CUC1), PRODUCTION-OF-ANTHOCYANIN-PIGMENT1 (PAP1), and AtMYB23 genes, which are involved in different physiologic phenomena, namely response to hormone, differentiation, biosynthesis of a metabolite, and development, respectively (Borevitz et al., 2000; Chao et al., 1997; Kirik et al., 2001; Takada et al., 2000). We reported that the chimeric proteins that included the EAR motif acted as dominant repressors. They overcame the redundancy of transcription factors and effectively suppressed their target genes in transgenic Arabidopsis, with resultant loss-of-function phenotypes.
The ethylene-insensitive phenotype of transgenic plants that expressed the chimeric EIN3 repressor
The EIN3 transcription factor is a positive regulator of ethylene-mediated effects (Chao et al., 1997). To construct the chimeric EIN3 repressor, we fused the coding region of EIN3 with the DNA encoding for the EAR-motif repression domain (RD) from tobacco ERF3 (Ohta et al., 2001) or from SUP (SUPRD), which exhibited stronger repressive activity than RD in transient expression assays (Hiratsu et al., 2002). The sequences were fused in frame, and the resultant polynucleotides were fused with the 35S promoter of cauliflower mosaic virus (CaMV) to yield 35S::EIN3RD and 35S::EIN3SUPRD (Figure 1a). In the 20 independent T1 transgenic lines that we examined, all 35S::EIN3RD and 35S::EIN3SUPRD plants failed to respond to ethylene, and the normal effects of ethylene, such as the triple response in etiolated seedlings and the inhibition of cell growth in rosettes, were not observed (Figure 1b). In addition, the ethylene-responsive expression of the ETHYLENE-RESPONSE-FACTOR1 gene, the PDF1.2 gene for defensin, and the gene for basic-chitinase, all of which are normally regulated by EIN3 (Chao et al., 1997; Solano et al., 1998), were also undetectable (Figure 1c). These results indicated that the transgenic Arabidopsis plants that expressed the chimeric EIN3 repressor were all insensitive to ethylene, resembling plants with loss-of-function alleles of EIN3 (Chao et al., 1997). By contrast, no ethylene-insensitive phenotype was observed in the case of transgenic plants that expressed the 35S::EIN3RDm transgene, in which EIN3 had been fused to a mutant version of repression domain (RDm; Figure 1), which lacked the ability to repress gene expression (Ohta et al., 2001; Figure 1b,c).
Expression of the chimeric CUC1 repressor resulted in cup-shaped cotyledons
The products of the CUC1 gene and its homolog CUC2 are functionally redundant transcriptional activators (Aida et al., 1997; Takada et al., 2000). In cuc1/cuc2 double-mutant plants, a defect in the separation of cotyledons results in the formation of a cup-shaped structure, while individual single-mutant plants are basically normal (Aida et al., 1997; Takada et al., 2000). We fused RD and SUPRD separately to CUC1 to generate chimeric CUC1 repressors, but most of the T1 transgenic plants that we obtained failed to display a loss-of-function phenotype, namely cup-shaped cotyledons (data not shown). We suspected that the repressive activities of RD or SUPRD might have been insufficient to overcome the activity of the functionally redundant transcription factor, CUC2, in transgenic plants. To overcome this putative redundancy, we maximized the potential repressive activity of the EAR-motif repression domain by minimizing the length and modifying the sequence of SUPRD (manuscript in preparation). We used the resultant repression domain, designated SRDX (LDLDLELRLGFA; Figure 2a), for the construction of chimeric repressors. When the chimeric protein in which CUC1 was fused to SRDX (35S::CUC1SRDX) was expressed, all the seedlings of T1 transgenic plants had fused cotyledons with two cotyledonary blades fused together, from the base, to varying extents (Figure 2b). Among the 63 T1 seedlings that we investigated, 16 had cup-shaped cotyledons similar to those in the cuc1/cuc2 double mutant (Aida et al., 1997). Most of the seedlings of 35S::CUC1SRDX plants failed to form a shoot apical meristem, as also reported in the cuc1/cuc2 double-mutant plants (Aida et al., 1997). The chimeric protein in which CUC2 was fused to SRDX (35S::CUC2SRDX) also caused the formation of cup-shaped cotyledons, similar to those observed in seedlings of 35S::CUC1SRDX plants (data not shown). Our results indicated that the chimeric protein with the modified EAR motif acted in a dominant manner in the presence of functionally redundant transcription factors.
Expression of the chimeric PAP1 repressor resulted in reduced accumulation of purple pigment and suppression of the expression of genes related to phenylpropanoid biosynthesis
The MYB family of plant transcription factors includes PAP1, which is a positive regulator of phenylpropanoid biosynthesis (Borevitz et al., 2000). When wild-type Arabidopsis plants are grown under stress (in MS medium that contains 3% sucrose), seedlings develop purple pigmentation that is characteristic of anthocyanins (Martin et al., 2002;Figure 3a). By contrast, in 20 independent T1 transgenic lines that expressed PAP1 fused to SRDX (35S::PAP1SRDX), all the seedlings exhibited minimal purple pigmentation, or none at all, in the presence of 3% sucrose (Figure 3a). We examined the expression of genes related to phenylpropanoid biosynthesis, namely genes for phenylalanine ammonia-lyase (PAL), chalcone synthetase (CHS), and dihydroflavonol reductase (DFR), because expression of these genes is enhanced in PAP1-overexpressing mutant plants (Borevitz et al., 2000). These genes were all expressed at elevated levels in wild-type plants in the presence of 3% sucrose, while no such enhanced expression of these genes was detected in 35S::PAP1SRDX plants grown under the same condition (Figure 3b). These results indicated that the chimeric PAP1 repressor was able to suppress the induction of the various biosynthetic genes, and thus to inhibit the accumulation of anthocyanin. Putative MYB-binding sites can be found in the 5′ upstream regions, both of the PAL and CHS genes (Feinbaum and Ausubel, 1988), suggesting that these genes are directly regulated by PAP1.
Expression of the chimeric AtMYB23 repressor resulted in reduced numbers of trichomes
The MYB transcription factor, AtMYB23, is a homolog of GLABRA1 (GL1) and a positive regulator of trichome development (Kirik et al., 2001; Oppenheimer et al., 1991). Transgenic plants that expressed the chimeric protein that consisted of AtMYB23 and SRDX (35S::AtMYB23SRDX) were characterized by the disruption to trichome development (Figure 4a). Among plants of 35 independent T1 lines investigated, six of 35S::AtMYB23SRDX plants had no trichomes and the rest had considerably fewer trichomes than the wild type. We examined the expression of genes related to trichome development, namely GL1, GL2, and TRANSPARENT TESTA GLABRA1 (TTG1; Rerie et al., 1994; Walker et al., 1999), and we found that the expression of only GL2 was suppressed in the 35S::AtMYB23SRDX plants while GL1 and TTG1 were unaffected (Figure 4b). The 5′ upstream region of the GL2 gene includes putative MYB-binding sites (Szymanski et al., 1998). Thus, it is likely that the chimeric AtMYB23 repressor directly and dominantly suppressed the expression of GL2. The expression of GL2 was completely suppressed in trichome-free leaves, while it was detected at a low level in leaves with reduced numbers of trichomes (Figure 4b), suggesting a correlation between the extent of inhibition of trichome development and that of the expression of GL2. It seems likely that GL2 might regulate the development of trichomes directly.
Dominant negative phenotypes were not because of co-suppression
We confirmed the expression of endogenous EIN3, CUC1, PAP1 and AtMYB23 genes, and of each of the transgenes in the respective transgenic plants (Figures 1d, 2c, 3c and 4c). Moreover, CUC2, PAP2, and GL1 redundant homologs of genes for CUC1, PAP1, and AtMYB23, respectively, were also expressed in transgenic plants (Figures 2c, 3c and 4b). These observations indicate that the loss-of-function phenotypes and the suppression of expression of target genes in the transgenic plants were not because of co-suppression. In addition, it is unlikely that the phenotypes induced by the chimeric repressors were a result of non-specific negative effects, such as transcriptional squelching, because such loss-of-function phenotypes were not observed in plants that ectopically expressed the various transcription factors (Borevitz et al., 2000; Kirik et al., 2001; Takada et al., 2000) or in transgenic plants that expressed transcription factors fused to a mutant form of the EAR-motif repression domain. We also confirmed that the phenotypes generated by the chimeric repressors were inheritable as dominant traits, except in the case of meristemless 35S::CUC1SRDX plants (data not shown).
In this study, we demonstrated that four different transcription factors, when fused to the EAR-motif repression domain, can act as dominant repressors in transgenic plants. In the case of the EIN3 gene, a mutant phenotype can result from a single mutation (Chao et al., 1997). We found that the repression domains from ERF3 (RD) and SUP (SUPRD) effectively converted EIN3 into a dominant repressor and produced transgenic plants with an ethylene-insensitive phenotype (Figure 2). However, in the case of CUC1, which encodes a protein that is made functionally redundant by CUC2, the repressive activities of RD and SUPRD seemed to be insufficient to alter the phenotype in the presence of the functionally redundant homolog. In addition, AtMYB23 has been shown to be partially redundant with respect to GL1 (Kirik et al., 2001), and PAP1 is strongly homologous to PAP2 (Borevitz et al., 2000). Therefore, we developed and made use of a strong repression domain, SRDX, instead of RD and SUPRD, to generate the chimeric repressors from the transcription factors for which putative functionally redundant homologs exist.
In studies by others, a synthetic DNA-binding domain and plant transcription factors, to which the human Sid or the Drosophila En repression domain had been fused, were found to suppress expression of a target gene or to induce phenotypes similar to those induced by loss-of-function alleles for the transcription factors in transgenic Arabidopsis (Guan et al., 2002; Markel et al., 2002). However, the repressive activities of these chimeric transcription factors in plant cells have not been characterized yet. By contrast, we confirmed that all the chimeric repressors used in our experiments acted as transcriptional repressors in transient expression assays in Arabidopsis (data not shown). Moreover, no chimeric repressors, other than those associated with the EAR motif, have been shown to act in a dominant manner in the presence of functionally redundant transcription factors in stable transgenic plants. We demonstrated that the repressive activity of the chimeric repressors that included the EAR motif was able to overcome the activity of functionally redundant transcription factors to induce a dominant loss-of-function phenotype, using CUC1/CUC2 as a model. In addition, PAP1 plus PAP2 and AtMYB23 plus GL1 are likely to be functionally redundant, and the plants with antisense constructs directed against the PAP1 or the AtMYB23 gene have no visibly abnormal features (Borevitz et al., 2000; Kirik et al., 2001).
It is surprising that a peptide of only 12 amino acids (LDLDLELRLGFA) can function as a potent repression domain that can convert transcriptional activators into dominant repressors in plant cells. To our knowledge, no other short peptide has been shown to have such an ability in plants. The phenotypes induced by the chimeric repressors derived from EIN3 and CUC1 were similar to those induced by loss-of-functional alleles, namely ein3 and cuc1/cuc2, respectively, and the expression of the chimeric repressors that included PAP1 and AtMYB23 resulted in respective anticipated abnormal phenotypes (Borevitz et al., 2000; Kirik et al., 2001). Moreover, although PAP1 and AtMYB23, both, include a strongly conserved MYB domain, the 35S::PAP1SRDX transgene did not inhibit trichome development and the 35S::AtMYB23SRDX transgene did not inhibit the accumulation of anthocyanin. These results suggest that the chimeric repressors suppressed expression of the same genes as those regulated by the original transcription factors, and that the fused EAR-motif repression domain did not affect the ability of the transcription factors to recognize their target genes or to interact with accessory factors.
We used the CaMV 35S promoter to drive the chimeric genes and succeeded in obtaining dominant-negative phenotypes using four transcription factors. However, a strong promoter does not seem, always, to be necessary for suppression of target genes because a much lower ratio of effector plasmid, encoding the repressor, to activator also resulted in effective suppression of the expression of the reporter gene in transient expression assays in Arabidopsis (Fujimoto et al., 2000). This observation suggests that the native promoter might be able to drive a chimeric repressor for expression in a defined tissue at a defined period of development.
Our chimeric repressor silencing technology (CRES-T) using the EAR motif has several advantages over the existing silencing methods. The system is unusually simple. Upon fusion of the short peptide, we can convert a transcription factor of interest into a dominant repressor. The chimeric repressor can induce a dominant-negative phenotype with high efficiency. In the case of EIN3, 100% of transgenic plants were ethylene insensitive. In the case of redundant transcription factors, namely CUC1, PAP1 and AtMYB23, 25, 50 and 17% of transgenic plants exhibited a strong dominant-negative phenotype, respectively, while the rest exhibited more modest disruption of the wild-type phenotypes. Using EAR-motif chimeric repressors, we might be able to analyze the functions of transcription factors in cases where single-gene knockout or antisense lines do not display visibly abnormal phenotypes, most probably, because of structural and functional redundancy. Our chimeric repressor system might be applicable not only to diploid plants but also to amphidiploid crop and horticultural plants, in which gene-knockout techniques might be less effective, because the EAR motif is found in rice, petunia and wheat, as well as in tobacco and Arabidopsis (Ohta et al., 2001). Thus, our system might not only facilitate the rapid analysis of the biologic functions of redundant plant transcription factors but also be useful for the manipulation of plant traits in agricultural biotechnology.
Cloning and transformation
The protein-coding regions of the EIN3, PAP1 and AtMYB23 genes were amplified by PCR from an Arabidopsis cDNA library, and those of CUC1 and CUC2 were amplified from cDNAs that were kindly provided from Dr M. Tasaka, with the appropriate 5′ and 3′ primers. The coding regions of the repression domains of ERF3 (aa 191–225) and SUP (aa 175–204) and the mutant repression domain RDm were prepared as described previously (Hiratsu et al., 2002; Ohta et al., 2001). Both strands of DNA fragments that corresponded to SRDX (LDLDLELRLGFA) were synthesized with a TAA termination codon at the 3′ end. The resultant DNAs were cloned into the transformation vector pBIG-HYG (Becker, 1990) with the CaMV 35S promoter and Nos terminator. These constructs were used to transform Agrobacterium tumefaciens strain GV3101 and were introduced into wild-type Arabidopsis plants (Col-0) by standard methods (Bechtold et al., 1993). Transgenic plants were selected on a hygromycin-containing medium.
Northern blotting analysis was performed as described previously, using total RNA extracted from leaves (Fujimoto et al., 2000). For RT-PCR analysis, 50–75 ng of total RNA, which has been extracted from leaves or seedlings and then treated with DNase, was subjected to first-strand cDNA synthesis. RT-PCR was performed with gene-specific primers (Table S1) for 25–35 cycles. The gene for β-tubulin was used as an internal control. The products of PCR were diluted into 10- to 100-fold, separated on an agarose gel, and subjected to DNA blot analysis using the corresponding cDNAs as probes. Signals corresponding to mRNAs were detected with the ECL system for direct labeling and detection of nucleic acids (Amersham, NJ, USA).
Table S1. Primers used for RT-PCR.
The authors thank Prof. M. Tasaka (NAIST) for the cDNA clones of CUC1 and CUC2 and for seeds of the cuc1/cuc2 mutant plants; the Arabidopsis Biological Resource Center (Ohio State University, Columbus) for ein3-1 seeds; and K. Yamaguchi and A. Yamanaka for their skilled technical assistance.