A new rice zinc-finger protein binds to the O2S box of the a-amylase gene promoter Rihe Peng 1 1 , Quanhong Yao 1 , Aisheng Xiong 1 , Huiqin Fan 1 , Xian Li 1 , Youliang Peng 2 , Zong-Ming Cheng 3 and Yi Li 4 1 Shanghai Key Laboratory of Agricultural Genetic and Breeding, Agro-Biotechnology Research Center, Shanghai Academy of Agricultural Sciences, China; 2 Department of Plant Pathology, Chinese Agricultural University, Beijing, 2 China; 3 Department of Plant Sciences, University of Tennessee, Knoxville, USA; 4 Department of Plant Science, University of Connecticut, Storrs, USA 3 A putative transcription facto r, named RAMY, that binds to the 20-bp O2S sequences of the regulatory region of the Amy2 gene promoter has been identified using the yeast one- hybrid system from a rice library. The full length RAMY cDNA clone encodes a 218-amino acid protein and is homologous to the late embryogenesis-abundant protein (LEA5). In vitro mutagenesis and electrophoretic mobility shift a ssays confirmed that RAMY can bind with O2S spe- cifically through an unusual zinc finger with a CXCX 4 CX 2 H consensus sequence. Low levels of RAMY mRNAs were detected in rice leaves and roots by Northern blot hybrid- ization. The plant hormone gibberellin (GA) induces expression of bo th RAMY and Amy2 genes, as performed by Northern blot hybridization, 4 buttheincreaseinRAMY mRNA level occurs prior to that of the Amy2 mRNA level in the GA-treated aleurone tissues. These data suggest that RAMY may act as a trans-acting protein and is probably involved in the GA-induced expression of the r ice a-amylase gene. Keywords: rice zinc-induced protein; O2S box; yeast one- hybrid system. Cereal a-amylase genes have been one of the primary systems for exploring t he molecular mechanisms involved in hormone-regulated gene expression in plants 5 .Duringger- mination of cereal grains, the embryo releases GA to the aleurone layer, where it induces the transcription of a-amylase genes [1]. Functional analyses of a-amylase gene promoters using transient expression assays with reporter genes have shown both GA and abscisic acid (ABA) may interact with transcriptional regulatory proteins or transcription factors that bind to a short nuclear nucleotide sequence referred to as the GA response element (GARE) [2–4]. Lanahan et al. [5] has demonstrated that GARE mediates the hormonal control of transcription in the promoter of the low-PI gene, Amy32b. At least three other distinct regulatory elements have been found to be necessary for high-level a-amylase gene expression regulated by GA. A closely associated group of elements is composed of an opaque-2-like protein binding seq uence (O2S), a sequence e lement with an enriched pyrimidine nucleotide motif (the pyrimidine box), the GARE, and box I ( TATCCAT) [6]. Using quantitative transient expression assays, the most import- ant e lements h ave b een found to be GARE and O 2S; mutation or deletion of either GARE or O2S resulted in lower GA-induced transcription [4,5]. Rogers and Rogers [4] found that both GARE and O2S functioned only when positioned in one orientation with respect to each other and with respect to the TATA box, and when the distance between them was relatively short. In searching for factors that interact with the sequences and regulate a-amylase gene expression, a Myb protein, GAMyb (GA-responsive Myb protein), was isolated that may specifically bind to a portion of the G ARE box [7]. GAMyb is able t o activate the expression of a-amylase and other GA-regulated genes [8]. A zinc-finger protein has been identified by South- western screening with baits containing a GARE box and has been found to repress the expression of a-amylase and other g enes [9]. In Arabidopsis, at least three proteins, SPY, RGA and GAI, are thought to negatively regulate GA responses [10–12]. However, although the O2S box is another important element for controlling the level of transcription in a-amylase gene, little is known about the regulatory proteins or transcription factors that bind to the O2S sequence. Our interest in the mechanism of seed germination and development in rice p rompted us to search for the O2S binding protein. Using the yeast one- hybrid system, we screened the rice cDNA libraries using O2S-containing baits ATTGACTTGACCGTCATCGG from the low pI amy54 promoter [13]. We have isolated a cDNA clone, RAMY, which encodes a protein that contains a zinc-finger. Our experimental data indicate that RAMY protein binds specifically to the O2S elemen t. We have also determined the importance of the amino acids within the binding domain of RAMY protein and analyzed the time course for the induction of RAMY a nd a-amylase mRNA by GA. Correspondence to Q. Yao, Shanghai Key Laboratory of Agricultural Genetic and Breeding, Agro-Biotechnology Research Center, Shang- hai Academy of Agricultural Sciences, 2901 Beidi Road, China. Fax: + 86 021 62209988, Tel.: + 86 021 62209988, E-mail: pengrihe69@yahoo.com Abbreviations: GST, glutathione S-transferase; O2S, opaque-2-like protein binding sequence; GA, gibberellin; ABA, abscisic acid; GARE, GA response element. (Received 30 January 2004, revised 28 A pril 2004, accepted 19 May 2004) Eur. J. Biochem. 271, 2949–2955 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04221.x Materials and methods Strains and plasmids Saccharomyces cerevisiae strain EGY4 8 ( MATa,his3trp1 ura3-52 leu::pLeu2-LexAop6) was used as the host strain for yeast transformations. Escherichia coli strain MC8 (thi – , trp – ,ura – ,leu – ,his – ) was used to rescue cDNA library plasmids from yeast DNA preparations. The bait plasmid pLGD-265 UP1 is a URA-marked 6 E. coli–yeast shuttle plasmid carrying a lacZ reporter gene under the control of the CYC1 minimal promoter [14]. The rice cDNA library was constructed in pPC86 vector, which is a marked 7 yeast expression plasmid containing a GAL4 activating domain under the control of the yeast ADC1 promoter [15]. The cDNA was derived from poly(A) RNA i solated from 15-day-old rice (Oryza sativa L., genotype IR36) seedlings grown at 25 °C in a greenhouse. Yeast one-hybrid screen A yeast one-hybrid screen was performed to isolate genes encoding proteins that associated with the O2S box ATTGACTTGACCGTCATCGG in the Amy2 gene pro- moter [12]. To construct bait plasmid, the cis-element containing three copies of O2S was synthesized by PCR using two primers: Amyb1: 5¢-ACCCTCGAGGTCGA CGGTATCGATAAGCTTGATTGACTTGACCGTCA TCGGATTGACTTGACCGTCATCG-3¢,Amyb2:5¢-CA GGATCCATCACGACAGTCAGTGCCGATGACGG TCAAGTCAATCCGATG-3¢ (the PCR conditions were: 94 °C, 20 s; 58 °C30s;72°C, 30 s; 25 cycles). Following PCR, the 110-bp fragment isolated from PAGE was digested with restriction enzymes, and inserted into the XhoI/BamHI sites of the vector pLGD 8 -265 UP1. The lithium acetate protocol was used for yeast transfor- mation [16]. An overnight culture of yeast cells (0.5 mL) were inoculated into 50 mL fresh YPD medium and grown for 4 h. Yeast cells were centirfuged at 5000 g for 8 min, and the pellets washed once with 20 mL sterile distilled water, and then with 10 mL Tris/EDTA/LiAc (100 m M LiAc in Tris/ EDTA). Finally, the cell pe llet was resuspended in 0.5 mL Tris/EDTA/LiAc. An aliquot of 50 lL yeast cells, 1 lg plasmid DNA, 50 lg salmon carrier DNA and 300 lL Tris/ EDTA/LiAc containing 40% (w/v) 9 PEG3350 was added to each tube, and the mixtures were incubated for 30 min at 30 °C with shaking (150 r.p.m.). After heat shock at 42 °C for 15 min, the collected yeast cells were resuspended i n Tris/ EDTA buffer and plated on selective yeast media. The total DNA from each positive yeast clone was isolated according to the method of Robzyk and Kassir [17]; DNA cloning was performed using to standard procedures [18]. The DNA extracted from yeast cells was electroporated into E. coli strain MC8 [19]. The transformants were selected on M9 minimal medium containing a ll amino acids except tryptophan (M9-TRP), thus selecting pPC86-con- taining colonies. The plasmid DNA was re-introduced into a yeast reporter strain to confirm the b-galactosidase activity. Two oligonucleotide primers GAL4 5¢-GGA TGTTTAATACCACT-3¢ and TAD4 5¢-TTGATTG GAGACTTGACC-3¢ derived from the DNA sequence flanking the GAL4 activating domain and the ADC1 terminator region in the pPC86 vector, respectively, were used to amplify the rice cDNA inserted into vector pPC86. PCR was carried out for 30 cycles according to the following protocol: 94 °C for 30 s, 42 °C f or 45 s, and 72 °C for 3 min. The DNA sequence analysis of the recombinant plasmids was performed with the model 373 ABI automatic sequencer. Southern–Northern blot analyses Extraction of plant DNA and Southern blotting analysis were p erformed b ased on the m ethod of Bringloe [20]. Total RNA was prepared from rice (O. sativa L. IR36); leaves and roots were treated with 10 lMGA,and control seedlings were treated with water by the standard method 10 [18]. In order to examine the time course of the GA-induced RAMY and Amy2 gene expression, steady state levels of their transcripts in the GA-treated aleurone tissues were determined with Northern blot hybridization. De-embryo- nated rice half-seeds were de-husked, sterilized with 10% (v/v) commercial Clorox, and treated with 95% (v/v) ethanol for 30 s to remove the outer wax layer. After rinsing with water, the rice de-embryonated half-seeds (20 g) were submerged in seed buffer (20 m M calcium chloride, 20 m M sodium succinate, pH 5.2) in cell culture dishes. After 10 h of incubation in seed buffer, appropriate amounts of GA were added to a final concentration of 10 l M . Isolated aleurone tissues were incubated with GA for2,4,6,8,10,12or14h. RNA was separated by gel electrophoresis an d trans- ferred to a H ybond N + nylon m embrane ( Amersham- Pharmacia). Hybridization was performed at 65 °Cin 5· NaCl/Cit, 10% (w/v) dextran sulfate, 0.5% (w/v) SDS and 0.1 mgÆmL )1 denatured salmon sperm DNA. Filters were washed twice (each for 15 min) at 65 °Cin2· NaCl/ Cit, 0.1% SDS, and once in 0.1· NaCl/Cit, 0.1% SDS at 65 °C for 15 min. Production and analyses of GST–RAMY fusion protein The GST–RAMY fusion protein was generated by inserting RAMY gene open reading frame between the sites of BamHI and SacI in the vector pALEX [21]. The primers used were RAMYZ (5¢-AGGATCCATGGCT CTCGCTCTCTCCACC-3¢)andRAMYF(5¢-AGA GCTCAGTGGTGGTGGTGGTGGTGCACTCGGGT ACGTGGTGAAAC-3¢). Mutated RAMY polypeptides were produced by in vitro mutagenesis u sing the follow- ing primers: C182SZ (5¢-CGTGTGGGCTGGATGCT CTCCT-3¢), C182SF (5¢-GAGGAGAGCATCCAGCCC AC-3¢), C184SZ (5¢-GTGGGCTGGATGCTCTGCTCG TCTG-3¢), C184SF (5¢-GCAGACGAGCAGAGC ATC CAGC-3¢), C189SZ (5¢-CTGCTCGTGTGCTGGTTCTT CGTCCAC-3¢), C189SF (5¢-GAGGTGGACGAAGAAC CAGCACACG-3¢), H192AZ (5¢-TGCTGGTTCTTCGT GCACCTCTGCTGTAAC-3¢), and H192AF (5¢-GTGAG GCCCTGCTCGCTGTTACAGCAGAGGTGC-3¢). All mutant genes were cloned into a pUC18 vector and sequenced for confirmation. The resultant mutant RAMY genes were then inserted into the BamHI/SacIsitesofthe vector pALEX. 2950 R. Peng et al. (Eur. J. Biochem. 271) Ó FEBS 2004 The resulting GST–RA MY construct was used to trans- form E. coli BL21. Transformants were used to inoculate 50 mL cultures of LB/ampicillin and were grown overnight at 37 °C and harvested by centrifugation. Cell pellets were resuspended in 9 mL phosphate buffer (pH 7.4), 20 m M imidazole was added, and the bacteria were lysed by sonication. After centrif ugation at 8000 g 12 for 10 m in at 4 °C, the GST–RAMY fusion protein was isolated using HiTrap chelating columns according to t he manufacturer’s instructions (Amersham-Pharmacia). SDS/PAGE a nd Coomassie blue staining were used to determine the purity of the protein. Protein concentrations were determined using a protein assay kit (Bio-Rad). Mutant GST–RAMY fusion protein 13 was expressed and purified in the same way as for the GST–RAMY fusion protein. The mutant O2S box was synthesized by PCR using the primers mAmyb1 (5¢-ACCCTCGAGATTGAGCTAGCC GTCATCGGATTGAGCTAGCCGTCATCG-3¢)and mAmyb2 (5¢-CAGGATCCGTCAGTGCCGATGACG GCTAGCTCAATCCGATG-3¢) The PCR conditions were: 94 °C, 20 s; 58 °C30s;72°C, 30 s for 25 cycles. The core binding sequence CTTGA in the conserved O2S domain was replaced by GCTAG. A gel retardation assay was carried out as described by Jensen [22]. The XhoIand BamHI fragment containing three copies of the O2S box was labelled using a random primer method with the wild- type or mutant O2S motifs as competitors. One nanogram of labelled O2S DNA fragments, competitor DNA, 2.5 lg poly(dI–dC) and 10 lg GST–RAMY binding protein were mixed in 25 lL of DNA binding buffer [5 m M Hepes pH 7.5, 2 m M MgCl 2 ,0.2m M dithiothreitol, 1 m M CaCl 2 , 2% (w/v) g lycerol 14 ]. The mixture was incubated for 20 min at room temperature and loaded onto a 6% polyacrylamide gel. After migration, the gel was fixed in 5% (w/v) glycerol, 5% (w/v) methanol, and 5% (w/v) acetic acid 15 . Labeled DNA was then transferred to Whatman paper 16 ,driedand autoradiographed. Results Screening for rice cDNA encoding the O2S binding protein To isolate genes whose products bind to the O2S domain (ATTGACTTGACCGTCATCGG) in the Amy54 gene promoter, two plasmids were used in the yeast one-hybrid system. The plasmid pPC86 contained a rice cDNA library to express GAL4–cDNA fusion proteins, and plasmid pLGD 17 -265UP1wasusedasthebaitwithan insertion of three copies of the O2S domain at the 5¢ end of the CYC1 mini-promoter region. Following transfor- mation of the URA-marked plasmid pLGD-265 UP1 containing the O2S domain and the TRP-marked pPC86 plasmid carrying rice cDNA library 20 into the yeast cells and selection on selective medium for 2 days, a pproxi- mately 5 · 10 6 yeast transformants were overlaid onto SC-TRP-URA X-gal medium using nitrocellulose filters. In the first round of selection, 31 positive (blue) colonies were selected. To verify the true positive clones after the first round selection, total yeast DNA was extracted and transformed into the E. coli strain MC8. Following selection for the E. coli strain MC8, 14 individual transformants were positive (blue) on medium containing X-gal. Nucleotide sequence and predicted amino acid sequence of rice RAMY cDNA Using GAL4 and TAD4 primers, sequencing analyses of both strands of all 14 cDNA fragment 21 revealed that they contained identical, o verlapping sequences en coding the same protein. The complete nucleotide s equence encoded a predicted protein of 218 amino acids (Fig. 1). A search in a number o f random protein d atabases 22 revealed that there is n o significant homology b etween Fig. 1. 32 The rice RAMY cDNA sequences and the predicted product of its longest ORF (GenBank accession no. AY072712). The putative DNA binding domain is underlined, and the putative nuclear localization signal is double-underlined. RAMY 60 DGSSSSA AREVSWVPDPVTGHYRPSNFAGGRRRRPPRRPPRP 101 G.max LEAS 59 DTRDGSK AYSTDWAPDPVTGYYRPINHTPEIDPVELRHRLLR 100 N.tabacum LEA5 51 KWEESS KKTTSWVPDPVTGYYRPESHAKEIDAAELRQMLLN 91 G.hirsutum LEA5 49 AMKESSSSETRAYSSAWAPDPVTGYYRPENCGAEIDAAELREMLLN 94 V.radiata ARG2 52 KSGEEKVR- GGEKVSWVPDPVTGYYRPEN-TNEIDVADMRATVLG 94 A.thaliana ARG21 53 KGVEES TQKISWVPDPKTGYYRPETGSNEIDAAELRAALLN 203 H.vulgare G3 59 REAEKA AADSSWVPDPVTGHYRPANRSSGADPADLRAAHLG 100 Fig. 2. Alignment of the conserved domain of RAMY with some related proteins. RAMY is compared with the LEA5 p roteins from G. max (accession no. AAB38782), N. tabacum (accession no. AAC06242) and G. hirsutum (accession no. P46522), the ARG2 proteins from V. radiata (accession no. P32292) and A. thaliana (accession no. AAC19273), and with the G3 protein from H. vulgare (accession no. CAA55482). Ó FEBS 2004 A new zinc-finger protein from rice (Eur. J. Biochem. 271) 2951 RAMY with other proteins deposited in the databases. However, the N-terminal half of RAMY is homologous to the LEA5 proteins from Glycine max (GenBank acces- sion no. AAB38782), Nicotiana tabacum (accession no. AAC06242) and Gossypium hir sutum (accession no. P46522), the ARG2 protein encoded by cDNAs isolated from Vigna radiata (accession no. P32292) [23] and Arabid- opsis thaliana (accession no. AAC19273), and the G3 protein encoded by Hordeum vulgare (accession no. CAA55482) [24] (Fig. 2). The C-terminal half of RAMY contained a motif with Cys and H is residues similar to the zinc finger C3H 23 (Fig. 1, DNA binding domain). Using the 720-bp full-length RAMY cDNA as a probe for hybridization of rice genomic DNA that had been digested separately with HindIII, XbaI, and BamHI, we observed only one DNA band that hybridized with the probe from each of the digested DNA samples (Fig. 3). Because there are no HindIII, XbaI, and BamHI sites within the 720-bp cDNA, a single band observed in the Southern blot hybridization experiment suggests a single copy of the RAMY gene is present in the rice genome. RAMY contains a novel zinc finger Using the yeast one-hybrid system, the GAL4–RAMY fusion protein exhibited a strong transcriptional activation function in yeast cells. The GAL4–RAMY fusion protein 24 bound to the cis-element in bait pLGD 25 -265UP1, and induced transcription of the LacZ reporter gene. To examine whether RAMY protein binds to the O2S sequence directly and specifically, we performed an elec- trophoretic mobility shift assay experiment. As shown in Fig. 4A, the presence of the purified GST–RAMY protein resulted in a mobility shift of the 32 P-labelled O2S DNA Fig. 3. Southern blot analysis of Oryza sativa genomic DNA. Total genomic D NA (1 0 lg per lane) was digested with the restriction enzymes, HindIII (H), XbaI(X)andBamHI (B). The positive control (CK) was pPC86 (RAMY) DNA digested with SalI. Full-length RAMY cDNA was used as the probe. The m olecular marker (in k b) was kDNA digested with EcoRI and HindIII. Fig. 4. Characterization of the DNA binding affinity of RAMY recombinant protein to the Amy2/O2S sequences. (A) The [a- 32 P]dATP labelled O2S probes were incubated in the presence or absence of GST or GST–RAMY. Lane 1, purified GST; lanes 2–5, GST–RAMY. Binding is competed by the same length of fragment containing three copies of unlabelled Amy2/O2S sequence; lane 2, 50 · competitor DNA; lane 3, 10 · competitor DNA; lane 4, 5 · competitor DNA; lane 5, 1 · c ompetit or DNA. (B) Comparison of DNA b inding pref- erences of RAMY protein to the mutants (M) and the wild-type O2S (W). Lane 1, Purified GST; lane 2, GST–RAMY; lanes 3–6, EMSA by preincubating 25-fold (25 ·) or 100 fold (100 ·)excessamountsof unlabelled DNA fragments. F, free probe; S, shift probe. 2952 R. Peng et al. (Eur. J. Biochem. 271) Ó FEBS 2004 fragment on the gel, suggesting that the GST–RAMY protein binds directly to the O2S sequence. Because the purified GST alone did not produce such a mobility shift, the shift must be caused by RAMY protein. Furthermore, the unlabelled O2S D NA competed w ell with t he 32 P- labelled O2S DNA but nonspecific DNA or mutant O2S DNA did not (Fig. 4B), demonstrating that the binding of RAMY to the O2S is specific. We per formed a n in vitro mutagenesis experiment t o identify amino acid residues of RAMY protein that are important for the DNA binding. In the C-terminal region of RAMY, there are three Cys and one His residues. Figure 5 2626 shows that mutagenesis of Cys182 severely reduced binding activity of RAMY to the O2S DNA sequence. Furthermore, mutagenesis of Cys184, Cys189 or His192 abolished the binding o f the protein to the O2S sequence completely. RAMY mRNA accumulation in GA-treated tissues To determine whether the expression of the RAMY gene is induced by GA, and the possible relationship between the expression of the RAMY gene and the Amy2 gene, we performed a Northern blot hybridization e xperiment. As showninFig.6A,RAMY was expressed at low levels in leaves but almost not at all in roots 27 . However, in both roots and leaves, expression of RAMY was significantly induced by exogenous GA. We also used aleurone tissues to study the time course of the GA-mediated induction of RAMY and Amy2 expression. Figure 6B shows that RAMY mRNA was observed 4 h after the GA treatment, but an increase in Amy2 mRNA was seen only after 10 h of the GA treatment. The level of RAMY mRNA reached its maximum 10 h after the GA treatment and returned to its basal value within 14 h. In contrast, Amy2 mRNA reached a maximum at 14 h after the GA treatment (Fig. 6B). These data suggest that RAMY could act as a r egulatory o r transcription factor for the expression of a-amylase genes. Discussion We have cloned a gene, RAMY, from a rice cDNA library using O2S-contain ing baits with the yeast one-hybrid method that encodes a zinc-finger protein 28 . We believe that the O 2S-box binding domain is present in the protein. This conclusion is supported by the observations that the RAMY Fig. 5. DNA binding is mediated by the zinc finger domain. (A) SDS/ PAGE showin g the purity of the recombinant GST-fusion proteins (lanes 2–6). The GST-fusion proteins were over-expressed in E. coli BL21 (DE3), extracted under non denaturing conditions and purified by affinity chromatography. (B) G el m obility assay with the purified GST–RAMY fusion protein (lane 2) or GST-mutant RAMY fusion protein (lanes 3–6), showing RAMY binding to DNA a nd the effects of mutations within the RAMY DNA-binding domain on the ability of the purified fusion protein G ST–RAMY to bind a prob e contain ing three copies of the Amy2/O2S region. Fig. 6. RAMY transcript levels in a rice plantsubjectedtoGAtreat- ment. (A) GA induced RAMY mRNA accumulation in leaves and roots. L, leaves; L + GA, leaves treatedwithGA;R,roots;R+GA, roots treated with GA. (B) Induction of RAMY protein and a-amylase (GenBank accession no. A F411220) by GA. Aleurone tissu e was incubated with GA for 2, 4, 6, 8, 10, 12 or 14 h. Total RNA samples were loaded (20 lg per lane), fractionated on a 1.2% formamide/ agarose gel, probed with a 32 P-labelled RAMY probe. 18S, RNA probed with a 32 P-labelled 18S rRNA probe. a-amy , RNA pro bed with a 32 P-labelled a-amylase gene probe. Ó FEBS 2004 A new zinc-finger protein from rice (Eur. J. Biochem. 271) 2953 gene expressed in E. coli produces a protein that binds to the O2S-box specifically and that repeated yeast one-hybrid screening of the rice cDNA library with the O2S bait resulted only in isolation of the RAMY cDNA. The C-terminal end of RAMY has a motif containing Cys and His residues reminiscent of a novel C3H zinc finger motif CXCX 4 CX 2 H and is important in its binding activity to the O2S domain. Several plant zinc-finger proteins are of the cluster type with multiple repeated fingers separated by sequences of different length. The spacing between the fingers is the main element to determine the specificity of binding target sequence [25]. However, we only detected one zinc-finger domain in RAMY, perhaps because the different structure of the zinc finger motif in RAMY influences the binding domain. A putative nuclear localization signal [26,27] sequence is also found in RAMY (Arg90–Arg97), suggesting that RAMY may be transported into the nucleus through the nuclear pore complex using its own nuclear localization signal. Database searches revealed RAMY was homologous only to part of the LEA5 (late embryogenesis-abundant) proteins. LEA5 proteins display a hydrophobic N-terminal half and a hydrophilic C-terminal half [28]. This family of proteins is further characterized by a highly conserved motif of 12 amino acids with the consensus WAPDPVTGYYRP. RAMY can be grouped into the Lea5 29 familybasedonthe presence of the canonical 12 amino acid sequen ce motif (WAPDPVTGYYRP). Most LEA5-like p roteins a re induced in embryos o r vegetative tissues by desiccation, ABA or high osmoticum. The soybean Lea5-like (D-73 like) cDNA accumulates in desiccating seeds from 20 to 80 days after flowering 30 and in roots but not in leaves of drought-stressed plants [29,30]. The related gene (Di21)fromArabidopsis displays increased transcript accumulation in roots and leaves after drought induction, but is not detected in mature dry seeds of nonstressed Arabidopsis plants [31,32]. In cotton, the Lea5 transcripts are highly induced in mature leave of water- stressed plants or in w ater-stressed detached leaves [28]. However, the evidence presented here indicates that RAMY transcription is induced by GA. The interaction between GA and ABA may be important in controlling the a-amylase gene expression. TheeffectsofGAandABAona-amylase gene transcription have been established from transient expres- sion experiments using a-amylase promoter-reporter gene constructs in aleurone protoplasts [33,34]. Two physically associated elements are essential: a GA response element (GARE) regulated by GA and ABA, and an opaque-2 binding sequence (O2S). This is consistent with the hypo- thesis that protein b inding and interaction between two separate binding sites are required for high-level transcrip- tion and proper hormonal regulation [5]. The expression of GAMyb at the mRNA level is upregulated by GA [7] and theincreaseintheGAMyb mRNA occurs before that of Amy21 mRNA after GA treatment. In addition, GAMyb specifically binds to an Amy21 GARE, and transient expression experiments have shown that GAMyb activates transcription of a high-pI a-amylase promoter fused to a reporter gene in the absence of GA. These results suggest that GAMyb is a GA-regulated transcription factor required for transcriptional activation of the high-pI a-amylase promoter [7]. Similarly, RAMY binds to the O2S element specifically and RAMY mRNA a lso accumu- lates prior to the accumulation of the Amy2 mRNA upon GA treatment. Although, our experimental results suggest that RAMY is probably a transcription factor for activation of a-amylase genes, the role of RAMY in t he GA-response pathway leading to a-amylase gene expression needs to be established. Acknowledgements We would like to thank Dr Qun Zhu for proving us the r ice cDNA library. This research was supported by China PR Committee of Science. References 1. Jacobsen, J.V., Gubler, F. & Chandler, P.M. (1995) Gibberellin action in germinated cereal grains. In Plant Hormones: Physiology, Biochemistry and Molecular B iology (Davies, P .J., ed.), pp. 246– 271. Kluwer Academic Publishers, Dordrecht, The Netherlands. 2. Skriver, K., Olsen, F.L., Rogers, J.C. & Mundy, J., (1991) cis- Acting DNA elements r esponsive to gibberellin and its antagonist abscisic acid. Proc. Natl Acad. Sci. USA 88, 7266–72770. 3. Gubler, F. & Jacobsen, J.V. (1992) Gibberellin-responsive ele- ments in the promoter of a barley high-pI a-amylase gene. Plant Cell 31 4, 1435–1441. 4. Rogers, J.C. & Rogers, S.W. (1992) Definition and functional implications of gibberellin and abscisic acid cis-acting hormone response complex. Plant Cell. 4, 1443–1451. 5. Lanahan, M.B., Ho, T H.D., Rogers, W.S. & Rogers, J.C. (1992) A gibberellin response complex in cereal a-amylase gene pro- moters. Plant Cell. 4, 203–211. 6. Huang, N., Sutliff, T.D., Litts, J.C. & Rodriguez, R.L. (1990) Classification and characterization of the rice a-amylase multi- gene family. Plant Mol Biol. 14, 655–668. 7. Gubler, F., Kalla, R., Roberts, J.K. & Jacobsen, J.V. (1995) Gibberellin-regulated expression of a myb gene in barley aleurone cells: evidence for Myb t ransactivation of a high pI a-amylase gen e promoter. Plant Cell. 7, 1879–1891. 8.Gubler,F.,Raventos,D.,Key,M.,Watts,R.,Mundy,J.& Jacobsen, J.V. (1999) Target genes and regulatory domains of the GAMYB transcriptional activato r in cereal aleurone. Plant J. 17, 1–9. 9. Raventos, D., Skriver, K., Schlein, M., Karnahl, K. & Rogers, S.W. (1998) HRT, a novel zinc finger, transcriptional repressor from barley. J. Biol. Chem. 273, 23313–23320. 10. Jacobsen, S.E., Binkowski, K.A. & Olszewski, N.E. (1996) SPINDLY, a tetratricopeptide repeat protein involved in gibberellin signal transduction in Arabidopsis. Proc. Natl Acad. Sci. USA 93, 9292–9296. 11. Peng, J., Carol, P., Richards, D.E., King, K.E., Cowling, R.J., Murphy, G.P. & Harberd, N.P. (1997) The Arabidopsis GAI gene defines a signaling pathway that negatively regulates gibberellin responses. Genes Dev. 11, 3194–3205. 12. Silverstone, A.L., Ciampaglio, C .N. & Sun, T. (1998) The Arabi- dopsis RGA gene encodes a transcriptional regulator re pressing the gibberellin signal transduction pathway. Plant Cell. 10, 155–169. 13. Huttly, A.K. & Baulcombe, D.C. (1989) A wheat a-Amy2 pro- moter is regulated by gibberellin in transformed oat aleurone protoplasts. EMBO J. 8, 1907–1913. 14. Guarente, L. (1983) Yeast promoters and lacZ fusions designed to study expression of cloned genes in yeast. Methods Enzymol. 101, 181–191. 2954 R. Peng et al. (Eur. J. Biochem. 271) Ó FEBS 2004 15. Chevray, E.M. & Nathans, D. (1992) Protein interaction clon- ing in yeast: identification of ma mmalian proteins that react with the leucine zipper o f Jun. Proc. Natl A cad. Sci. USA 89, 5789– 5793. 16. Gietz, D., Jean, A.S. & Woods, R.A. (1992) Improved me thod for high efficiency transformation of intact yeast cells. Nucleic Acids Res. 20, 1425. 17. Robzyk, K. & Kassir, Y. (1992) A simple and highly efficient procedure for rescuing autonomous plasmi ds from yeast. Nucleic Acids Res. 20, 3790. 18. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Pr ess, New Yo rk, USA. 19. Dower, S.J., Miller, J.F. & Ragsdale, C.W. (1988) High efficiency transformation of E. coli by high voltage electropo ration. Nucleic Acids Res. 16, 6127–6145. 20. Bringloe, D.H., Dyer, T.A. & Gray, J.C. (1995) Developmental, circadian and light regulation of wheat ferredoxin gene expression. Plant Mol. Biol. 27, 293–306. 21. Panagiotidis, C.A. & Silverstein, S.J. (1995) pALEX, a dual-tag prokaryotic expression vector fo r the purification of full-length proteins. Gene. 164, 45–47. 22. Jensen, E.O., Marcker, K.A., Schell, J. & de Bruijin, F.J. (1988) Interaction of a nodule specific, trans-actin g factor with distinct DNA elements in soybean leghaemoglobin Ibc35¢ upstream region. EMBO J. 7, 1265–1271. 23. Yamamoto, K.T., Mori, H. & Imadeki, H. (1992) Novel mRNA sequences inducted by i ndole-3-acetic acid in se ctions of elongat- ing hypocotyls of mung bean (Vigna radiata). Plant Cell Physiol. 33, 13–20. 24. Speulman, E. & Salamini, F. (1995)£ÓGA3-regulated cDNAs from Hordeum vulgare leaves. Plant Mol Biol. 28, 915–926. 25. Meshi, T. & Iwabuchi, M. (1995) Plant transcription factors. Plant Cell Physiol. 36, 1405–1420. 26. Park, K.J. & Kanehisa, M. (1998) NLS (nuclear localization sig- nal) prediction. ICR Annual Report 5, 52–53. 27. Robbins, J., Dilworth, S.M., Laskey, R.A. & Dingwall, C. (1991) Two interdependent basic domains i n nucleoplasmin nuclear tar- geting sequence: identification of a class of bipartite nuclear tar- geting sequence. Cell 64, 615–623. 28. G 33 alau, G.A., Wang, H.Y C. & Hughes, D.W. (1993) Cotton Lea5 and Lea14 encode atypical late embryogenesis-abun dant proteins. Plant Physiol. 101, 695–696. 29. Yamamoto, K.T. (1994) Further characterization of auxin-regu- lated mRNAs in hypocotyl sections of mung bean Vigna radiata (L.) Wilczek: sequence homology to genes for fatty- acid desaturases a nd atypical late-embryog enesis- abundant pro- tein, and the mode of expression of the mRNAs. Planta 192, 359–364. 30. Burns, W.C., Maitra, N. & Cushman, J.C. (1996) Isolation a nd characterization of a cDNA enc oding a LEA5-like protein fro m soybean (U66316) (PGR96-103). Plant Physiol. 112, 1398. 31. Gosti,F.,Bertauche,Vartanian,N.&Giraudat,J.(1995)Abscisic acid-dependen t and -ind ependen t regula tion of g ene exp ression by progressive drought in Arabidopsis thaliana. Mol. Gen. Genet. 246, 10–18. 32. Xu,D.,Duan,X.,Wang,B.,Hong, B., Ho , T H.D. & Wu, R. (1996) Expression of a late embryogenesis abundant protein gene, HVA1, from barley confers tolerance to water deficit and salt stress in transgenic rice. Plant Physiol. 110 , 249–257. 33. Bethke, P.C., Schuurink, R. & Jones, R.L. (1997) Hormonal sig- naling in cereal aleurone. J. Exp Bot. 48, 1337–1356. 34. Love gro ve, A. & Hooley, R. (2000) Gibberellin and abscisic a cid signaling in aleurone. Trends Plant Sci. 5, 102–110. Ó FEBS 2004 A new zinc-finger protein from rice (Eur. J. Biochem. 271) 2955 . Similarly, RAMY binds to the O2S element specifically and RAMY mRNA a lso accumu- lates prior to the accumulation of the Amy2 mRNA upon GA treatment. Although,. a GARE box and has been found to repress the expression of a- amylase and other g enes [9]. In Arabidopsis, at least three proteins, SPY, RGA and GAI, are