FUSCA3 from barley unveils a common transcriptional regulation of seed-specific genes between cereals and Arabidopsis


  • Miguel Ángel Moreno-Risueno,

    1. Laboratorio de Bioquímica y Biología Molecular, Departamento de Biotecnología, Centro de Biotecnología y Genómica de Plantas – Universidad Politécnica de Madrid, ETS Ingenieros Agrónomos, 28 040 Madrid, Spain
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Noemi González,

    1. Laboratorio de Bioquímica y Biología Molecular, Departamento de Biotecnología, Centro de Biotecnología y Genómica de Plantas – Universidad Politécnica de Madrid, ETS Ingenieros Agrónomos, 28 040 Madrid, Spain
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Isabel Díaz,

    1. Laboratorio de Bioquímica y Biología Molecular, Departamento de Biotecnología, Centro de Biotecnología y Genómica de Plantas – Universidad Politécnica de Madrid, ETS Ingenieros Agrónomos, 28 040 Madrid, Spain
    Search for more papers by this author
  • François Parcy,

    1. Laboratoire de Physiologie Cellulaire Vegétale Centre National de la Recherche Scientifique (CNRS), Commissariat à l’Energie Atomique (CEA), Institut National de la Recherche Agronomique (INRA), Université Joseph Fourier, 17 rue des Martyrs, Batiment C2-38 054, Grenoble Cedex 9, France
    Search for more papers by this author
  • Pilar Carbonero,

    1. Laboratorio de Bioquímica y Biología Molecular, Departamento de Biotecnología, Centro de Biotecnología y Genómica de Plantas – Universidad Politécnica de Madrid, ETS Ingenieros Agrónomos, 28 040 Madrid, Spain
    Search for more papers by this author
  • Jesús Vicente-Carbajosa

    Corresponding author
    1. Laboratorio de Bioquímica y Biología Molecular, Departamento de Biotecnología, Centro de Biotecnología y Genómica de Plantas – Universidad Politécnica de Madrid, ETS Ingenieros Agrónomos, 28 040 Madrid, Spain
      (fax +34 91 336 39 85; e-mail jesus.vicente@upm.es).
    Search for more papers by this author

(fax +34 91 336 39 85; e-mail jesus.vicente@upm.es).


Accumulation of storage compounds in the embryo and endosperm of developing seeds is a highly regulated process that allows seedling growth upon germination until photosynthetic capacity is acquired. A critical regulatory element in the promoters of seed storage protein (SSP) genes from dicotyledonous species is the RY box, a target of B3-type transcription factors. However, the functionality of this motif in the transcriptional regulation of SSP genes from cereals has not been fully established. We report here the identification and molecular characterization of barley FUSCA3, a B3-type transcription factor as yet uncharacterized in monocotyledonous plants. Our results show that both the barley and Arabidopsis FUS3 genes maintain a conserved functionality for the regulation of SSP genes and anthocyanin biosynthesis in these two distantly related phylogenetic groups. Complementation of the loss-of-function mutant fus3 in Arabidopsis by the barley HvFus3 gene resulted in restored transcription from the At2S3 gene promoter and normal accumulation of anthocyanins in the seed. In barley, HvFUS3 participates in transcriptional activation of the endosperm-specific genes Hor2 and Itr1. HvFUS3, which specifically binds to RY boxes in EMSA experiments, trans-activates Hor2 and Itr1 promoters containing intact RY boxes in transient expression assays in developing endosperms. Mutations in the RY boxes abolished the HvFUS3-mediated trans-activation. HvFus3 transcripts accumulate in the endosperm and in the embryo of developing seeds, peaking at mid maturation phase. Remarkably, HvFUS3 interacts with the Opaque2-like bZIP factor BLZ2 in yeast, and this interaction is essential for full trans-activation of the seed-specific genes in planta.


In the evolutionary history of plants, seed formation is a key step that favours dispersal and makes possible interruption of the life cycle, enabling the plant to withstand adverse conditions in a quiescent state. Under a newly favourable environment, seeds can resume growth using storage compounds that supply nutrients to the embryo until photosynthetic capability is acquired. During seed formation, reserves are stored throughout the maturation phase, a period that is characterized by growth arrest, acquisition of desiccation tolerance and subsequent entry into a quiescent phase of variable length (Harada, 1997; Vicente-Carbajosa and Carbonero, 2005). In Arabidopsis thaliana, essential transcription factors (TFs) that participate in this process have been characterized, such as FUSCA3 (FUS3), ABSCISIC ACID-INSENSITIVE3 (ABI3), LEAFY COTYLEDON1 (LEC1) and LEC2, which genetically control various aspects of the maturation phase. FUS3, ABI3 and LEC2 belong to the VP1/ABI3-like B3 family of TFs, whereas LEC1 is a CBF-A protein (Giraudat et al., 1992; Lotan et al., 1998; Luerssen et al., 1998; Stone et al., 2001). Loss-of-function mutations in any of these genes severely affect the establishment of seed traits, leading to reduced expression of seed storage proteins (in fus3, abi3, lec1 and lec2), absence of chlorophyll degradation (in abi3) and of anthocyanin accumulation (in fus3, lec1 and lec2), reduced sensitivity to abscisic acid (in abi3 and lec1), intolerance to desiccation (in fus3, abi3 and lec1) and altered cotyledon identity (in fus3, lec1 and lec2) (Bäumlein et al., 1994; Keith et al., 1994; Kroj et al., 2003; Lotan et al., 1998; Meinke, 1992; Meinke et al.,1994; Nambara et al., 1995; Parcy et al., 1994, 1997; Raz et al., 2001; To et al., 2006; West et al., 1994). In addition, these regulators are interlocked in a complex hierarchical network of mutual interactions (To et al., 2006).

In dicotyledonous plants, cotyledons are the main tissue for reserve accumulation. In Arabidopsis, only a single layer of endosperm remains in the dry seed, and storage compounds are deposited in the embryo itself (Berger, 2003; Higgins, 1984; Vicente-Carbajosa and Carbonero, 2005). SSP genes are under tight tissue-specific and temporal transcriptional control during the maturation phase. In Brassica napus, precise cis regulatory elements (CREs) conferring seed-specific expression have been identified in the NapA gene promoter, including the B-box comprising the DistB (5′-GCCACTTGTC-3′) and ProxB (5′-CAAACACC-3′) elements, the RY box (5′-CATGCA-3′) and the G-box (5′-CACGTG-3′; Ezcurra et al., 1999). These CREs are also conserved and functionally relevant in the promoters of genes encoding 2S albumins from A. thaliana (Kroj et al., 2003). The importance of RY boxes and B3-type TFs in such regulation has been extensively documented. FUS3 and LEC2 seem to regulate SSP genes by directly binding to the RY box in their promoters. In particular, FUS3 specifically binds in vitro to the RY box in the At2S3 and NapA gene promoters (Kroj et al., 2003; Reidt et al., 2000). However, FUS3 has also been reported to regulate the synthesis of SSPs through TRANSPARENT TESTA GLABRA 1 (TTG1) or some other intermediate protein (Gazzarrini et al., 2004; Tsuchiya et al., 2004). Although ABI3 regulation of SSP genes is also dependent on the RY box, it is as yet unclear whether it takes place through a direct interaction or by activation of downstream targets such as FUS3 or LEC2. Moreover, a protein complex of ABI3 and two OPAQUE2-like (O2-like) bZIP factors, AtbZIP10 and AtbZIP25, that specifically bind to the G-box of the At2S1 gene promoter, has been reported to synergistically activate its expression (Lara et al., 2003).

In contrast to dicotyledonous plants, the endosperm is the prevailing reserve tissue in which starch and SSPs accumulate in the majority of monocot seeds. Regulation of SSP genes has been intensively studied, and numerous important cis-elements have been identified. In the promoters of genes encoding cereal SSPs, a bipartite endosperm box encompassing the GCN4 like-motif (GLM, 5′-ATGAGTCAT-3′) and the prolamin box (PB, 5′-TGTAAAG-3′), together with the AACA motif, are important regulatory elements (Forde et al., 1985; Takaiwa et al., 1996; Wu et al., 2000). These CREs are bound by TFs of the bZIP, DOF (DNA binding with one finger) and R2R3MYB families, respectively. In barley, the O2-like bZIPs, BLZ1 and BLZ2, and the DOF proteins, BPBF and SAD, activate expression of the Hor2 gene encoding the B-hordein storage protein, through these cis-elements (Díaz et al., 2005; Mena et al., 1998; Oñate et al., 1999; Vicente-Carbajosa et al., 1998). GAMYB, a member of the R2R3MYB family, has been reported to activate the expression of endosperm-specific genes such as the Itr1 gene, which encodes the trypsin inhibitor BTI-CMe, through the 5′-(C/T)AACA-3′ motif (Díaz et al., 2002). In addition, HvMCB1 and HvMYBS3, which belong to the R1MYB-SHAQYF class, regulate endosperm-specific genes through the 5′-GATA-3′ element (Rubio-Somoza et al., 2006a,b). Transcriptional regulation of SSPs and other endosperm-specific genes is achieved by the combinatorial effect of various TFs.

Most interestingly, the prolamins (the main group of SSPs from cereals) and a major group of dicot seed albumins can be phylogenetically traced to a common ancestor (Kreis and Shewry, 1989). Their conservation relates not only to the protein structure, but also to regulatory elements that are functionally exchangeable between the groups (Vincentz et al., 1997). In maize, VIVIPAROUS1, which is the ABI3 orthologue, is also involved in abscisic acid signalling, establishment of dormancy, and activation of maturation-specific genes during seed development (McCarty et al., 1989, 1991; Suzuki et al., 2003). Natively expressed in the embryo and aleurone, maize VP1 can complement Arabidopsis abi3 mutants that are impaired in the expression of SSP genes (Suzuki et al., 2001).

In support of the evolutionary relationship between prolamins and the dicot SSPs, the RY box is also conserved in the promoters of SSP genes from cereals, but has not been reported to play a role in their transcriptional regulation or to directly interact with a B3-type TF. Given that cereal prolamins are expressed in the endosperm and not in the embryo, and that VP1 is practically absent from the endosperm (McCarty et al., 1989), we explored the possibility that they could be under the regulation of a FUS3-like protein that, in Arabidopsis, is expressed both in the embryo and in the endosperm. Here, we report the identification and characterization of barley FUSCA3 (HvFUS3, encoded by the HvFus3 gene, accession number AM418838), and show that it specifically binds to and activates the Hor2 and Itr1 genes through the RY box in their promoters. Moreover, we have determined that HvFUS3 interacts in vivo with the O2-like bZIP factor BLZ2, and that both contribute to transcriptional activation of these two seed-specific genes. Complementation of the A. thaliana loss-of-function mutant fus3 by HvFus3 resulted in SSP gene activation and restoration of anthocyanin accumulation to wild-type levels in the Arabidopsis seed.


Identification of conserved putative VP1/ABI3-like B3 binding sites in the promoters of cereal seed-specific genes

As a first step in a systematic search for the FUSCA3 transcription factor (TF) from barley, we explored whether the putative cis-motif recognized by this TF (the RY box; Reidt et al., 2000) is present in the promoters of cereal endosperm-specific genes. We analysed the promoters of the barley Hor2 and Itr1 genes, which encode a B-hordein storage protein and the trypsin inhibitor BTI-CMe, respectively (Díaz et al., 1995, 2005; Rodriguez-Palenzuela et al., 1989). As shown in Figure 1(a), we found putative RY boxes in both promoters within the first 300 bp upstream of the translation start codon. Extending this analysis to wheat, rice, maize and Coix SSP gene sequences showed that putative RY boxes appear to be conserved in many of the studied promoters from these species: α-gliadins and high-molecular-weight glutenins from wheat, type A and B rice glutenins, maize 22 and 15 kDa zeins, and Coixα-coixin genes. This observation supports a putative role for the RY box in endosperm-specific gene expression, and prompted us to search for the orthologue of the Fus3 gene in one monocotyledonous species, namely barley.

Figure 1.

 Conservation of the RY box in the promoters of cereal endosperm-specific genes and phylogenetic relationships of the Arabidopsis, barley and rice Fus3 putative orthologues.
(a) Promoter sequence alignment of barley genes encoding seed storage proteins and trypsin inhibitor BTI-CMe in the region of the RY box element. The location of RY box sequences is given as the number of nucleotides upstream of the translation start codon. The genes are Hor2 (X87232), Itr1 (X65875), α-gliadin (K03076), high-molecular-weight glutelin (X12929), GluB glutelin (X54193), GluA glutelin (Os01g55690), 22 kDa zein (X55722), 15 kDa zein (M13507) and α-coixin (X63113) (accession numbers in parentheses).
(b) Unrooted phylogenetic tree of the Arabidopsis VP1/ABI3-like B3 genes [At3g26790 (AtFus3), At1g28300 (AtLec2), At3g24650 (AtAbi3), At1g01030, At2g30470, At2g36080, At2g46870, At3g11580, At3g61970, At4g01500, At4g21550, At4g32010, At5g06250] and the barley (HvFus3, AM418838) and rice (OsFus3, Os01g51610) putative orthologues of the AtFus3 gene. The tree was constructed using the neighbour-joining method based on alignment of the full-length protein sequences. Bootstrapping values are indicated as percentages, and branches that are less related to AtFUS3 were collapsed.
(c) Identification and distribution of conserved motifs among the HvFus3, OsFus3 and AtFus3 putative orthologues. The motifs, located in their relative positions within the protein sequences, were identified using MEME software based on the deduced amino acid sequences of the proteins encoded by the genes shown in (a).

Cloning of the barley HvFus3 gene

We initially searched barley EST databases for genes with homology to AtFUS3 (gene At3g26790), but failed to identify any sequence homologous to the Arabidopsis gene. We extended the search to the rice databases, and found that the rice locus Os01g51610 matched the sequence of the Arabidopsis Fus3 gene with an e-value of 4.0 e−31. The sequence of this gene, which we designated OsFus3, was used in a new search for barley ESTs, resulting in the identification of clones BM376703 and BM376595, with significant homologies. Based on their sequences, we amplified the full-length cDNA of the barley Fus3 gene (HvFus3) from a barley (cv. Bomi) developing endosperm sample.

We performed a phylogenetic analysis on FUS3- and ABI3-related genes from several species. The un-rooted phylogenetic tree obtained (Figure 1b) showed that the genes HvFus3, OsFus3 and AtFUS3 are phylogenetically related with a bootstrap value of 78 (out of 100), indicating that they are orthologous genes in barley, rice and Arabidopsis. Protein sequence analysis using MEME software identified seven conserved motifs in the sequences of HvFus3, OsFus3 and AtFUS3. The rice and barley FUS3 proteins shared motifs 1–6 and motifs 2–5, respectively, with the AtFUS3 protein. Motif 1 was only present in the cereals, and motif 7 was shared in AtFUS3 and the protein encoded by the At4g01500 gene.

In the Arabidopsis fus3 mutant background, HvFus3 activates transcription from the At2S3 gene promoter and restores anthocyanin accumulation to wild-type levels

To test the hypothesis of a conserved function between HvFus3 and AtFUS3, we transformed the Arabidopsis fus3 mutant with HvFus3, and, as a control, with AtFus3. Binary plasmids P35S:HvFus3 and P35S:AtFUS3 were transformed into the fus3-3 gl1 mutant containing the GFP gene under the control of the promoter of the At2S3 gene (accession number At4g27160; Kroj et al., 2003; Luerssen et al., 1998) that encodes the 2S albumin seed storage protein. Albumin 2S and its expression have been shown to be strongly reduced in fus3 mutant seeds. Seeds of wild-type and fus3-3 gl1 plants, both transformed with the GFP gene under the control of the At2S3 gene promoter, were analysed for green fluorescence and under bright-field microscopy. As shown in Figure 2, transcription of the GFP gene, which is strongly diminished in the fus3 mutant background, was restored to almost wild-type levels in seeds constitutively expressing HvFUS3 or AtFUS3. In contrast to wild-type seeds, fus3 mutant seeds are purple at the end of their development because anthocyanin biosynthesis is de-repressed in this mutant background (Figure 2) (Luerssen et al., 1998). In fus3-3 gl1 plants complemented with either the HvFus3 or AtFUS3 genes, we observed green-coloured seeds similar to those of the wild-type plants (Figure 2).

Figure 2.

 Complementation of the Arabidopsis fus3 mutant (Luerssen et al., 1998) with HvFus3 and AtFus3.
Wild-type and fus3 plants containing the GFP gene under the control of the At2S3 (2S albumin) gene promoter were transformed with the barley and Arabidopsis Fus3 genes under the control of the CaMV 35S promoter. Seeds of transformed plants were analysed for green fluorescence and under bright-field microscopy, to assess expression from the 2S3 promoter and anthocyanin accumulation in seeds, respectively. #1 and #2 correspond to independent transgenic lines.

Expression patterns of HvFus3 during barley seed development

Extensive homology searches using the HvFUS3 sequence in EST databases only produced positive matches with clones derived from seed tissues. The expression pattern of the HvFus3 gene in barley developing endosperms was investigated using real-time quantitative PCR. Endosperms were separated from the rest of the seed and collected on various days after pollination (10, 14, 18, 22 and 26 d.a.p.). In addition, embryos at 22 d.a.p. were also sampled. Variation in the transcript levels was analysed and normalized to those for the barley Actin2 gene.

As shown in Figure 3(a), the transcripts of HvFus3 increased by 1.5-fold between 10 and 22 d.a.p., then decreased at 26 d.a.p. to the same level detected at 10 d.a.p.. This expression pattern is similar to that of HvBlz2 (accession number Y10834, a known regulator of the Hor2 gene), which encodes a barley bZIP factor of the Opaque-2 class (Oñate et al., 1999). Therefore, the HvFus3 transcript expression profile is consistent with the possibility of this factor being a regulator of the Hor2 and Itr1 genes in the developing endosperm.

Figure 3.

 Expression analyses of the barley HvFus3gene by real-time quantitative PCR and in situ hybridization during barley seed maturation.
(a) Relative mRNA content of the HvFus3 and Blz2 TF genes and the Hor2 and Itr1 genes encoding the seed storage protein B-hordein and the trypsin inhibitor BTI-CMe, respectively, in developing endosperms from 10 to 26 d.a.p. All values were standardized to the barley Actin2 mRNA levels.
(b) Spatial expression pattern of the HvFus3 gene in developing barley seeds at 22 d.a.p. by in situ hybridization analysis. Hybridization signals were obtained using the antisense HvFus3 probe (upper and middle panels) or the sense probe (lower panel) as a negative control. a, aleurone; en, endosperm; er, embryo root; fp, foliar primordia; pe, pericarp; sc, scutellum; t, testa. Scale bars = 250 μm.

To determine the spatial expression of HvFus3 within developing seeds, in situ hybridization assays were performed in samples collected at 22 d.a.p.. In samples hybridized with the corresponding specific antisense probe, a strong signal was observed in endosperm and aleurone cells, as well as in the embryo, but not in the testa or the pericarp (Figure 3b, upper and middle panels). Signal intensities were in agreement with quantitative PCR values obtained for developing endosperm (approximately 1.0) and embryo (0.54). When samples were hybridized with the sense probes, used as negative controls, no signal above background was detected (lower panel).

HvFUS3 binds to the RY box in the promoters of the At2S3, Hor2 and Itr1 genes

As HvFUS3 activates transcription of the At2S3 promoter in Arabidopsis, and putative cis-motifs capable of being recognized by this TF are conserved in the promoters of many seed-specific genes, we tested whether HvFUS3 could specifically bind in vitro to the RY boxes present in the promoters of Hor2, Itr1 and At2S3 genes using electrophoretic mobility shift assays (EMSA).

The HvFUS3 protein and a mutated version in which the aspartic acid residue at position 75 was changed to asparagine (HvFUS3*), with native Arabidopsis FUS3 as a control, were expressed as GST fusions in Escherichia coli. The HvFUS3* mutation was designed on the basis of an abi3 mutant that contained this amino acid change at the equivalent residue of the B3 domain, which presumably impairs its DNA-binding capacity. As shown in Figure 4(a), when the three proteins were incubated with the radioactively labelled oligonucleotide probe At2S3-RY derived from the At2S3 gene promoter, a retarded band was observed for the HvFUS3 and AtFUS3 recombinant proteins but not for the mutated version HvFUS3*. Similarly, when we assayed the probe Hor2-RY derived from the Hor2 gene promoter, we detected a retardation band for the barley HvFUS3 protein but not for its mutated version. In addition, HvFUS3 also bound in vitro to the Itr1-RY probe derived from the Itr1 gene promoter (Figure 4b). Binding specificities were confirmed by competition titrations, using up to 50-fold molar excess, using non-labelled Itr1-RY and Itr1-ry probes in which the CATGCA core sequence was changed to ATCCTG.

Figure 4.

 Electrophoretic mobility shift assays (EMSA) of recombinant AtFUS3, HvFUS3 and mutated HvFUS3* proteins using probes derived from the At2S, Hor2 and Itr1 gene promoters.
(a) EMSA of the recombinant proteins AtFUS3, HvFUS3 and mutated HvFUS3* (aspartic acid at position 75 was changed to asparagine) using the 32P-labelled probes At2S3-RY and Hor2-RY, derived from the At2S3 and Hor2 gene promoters.
(b) EMSA of the HvFUS3 protein with 32P-labelled probes derived from the Itr1 gene promoter. Itr1-RY corresponds to the native sequence and Itr1-ry is a mutated version (lower case). Competition experiments were performed using increasing amounts (10 x, 25 x and 50 x) of the indicated unlabelled wild-type or mutated probes. The B3-binding target sequence is shown in bold.

HvFUS3 activates transcription of the Hor2 and Itr1 promoters in a RY box-dependent manner

To evaluate the functional relevance of the interaction observed in vitro between HvFUS3 and the RY box in the promoters of the Hor2 and Itr1 genes, we carried out transient expression assays in co-bombarded barley endosperms (cv. Bomi). Promoter:GUS reporters (Figure 5c), corresponding to constructs PHor2 and PItr1 (wt) as described by Díaz et al. (2002), and a newly generated construct PItr1 (mut), in which the native RY box (5′-CATGCA-3′) at position -147 was changed to 5′-ATCCTG-3′, unable to interact with HvFUS3 protein in vitro were assayed. Effector constructs included HvFus3 and two mutated versions of it: (i) HvFus3* (D75N, described above), and ii) HvFus3Δ7−46, which has a deletion of the 7th to the 46th amino acid residues. The ORF of HvBlz2 was also used as an effector. All these effector constructs were driven by the CaMV 35S promoter followed by the first intron of the maize alcohol dehydrogenase I (AdhI) gene, and flanked downstream by the 3′nos terminator. Freshly isolated developing barley endosperms were bombarded either with the reporters alone or in combination with the effectors at a 1:1 molar ratio. The GUS activity obtained after bombardment of the reporters without effectors was arbitrarily assigned a value of 100%. As shown in Figure 5(a), when we co-transfected PHor2 and PItr1 (wt) together with HvFUS3, the GUS activity increased threefold, but use of PItr1 (mut) did not result in any significant increase (Figure 5a). It is worth noting that this mutation in the RY box of the Itr1 promoter did not prevent its trans-activation by BLZ2, but reduced fivefold (inset to Figure 5a) the intrinsic expression of the native promoter, indicating the importance of this cis-motif. Increased activation was observed for combinations of HvFUS3 and BLZ2 on these endosperm-specific gene promoters. Thus, when both HvFUS3 and BLZ2 were co-bombarded as effectors with the reporters PHor2 and PItr1 (wt), higher GUS activities were obtained compared to individual effectors (Figure 5a), and this was especially conspicuous in the case of PHor2. However, when HvFUS3 and BLZ2 were co-bombarded using the mutated version PItr1 (mut) as the reporter, we did not observe any increase above the GUS activity detected when BLZ2 was used as the sole effector. Moreover, a mutation in either the B3 domain (FUS3*), abolishing DNA binding (Figure 4a), or the mutation FUS3Δ7−46 (with deletion of a conserved N-terminal region), which lacks activating capacity, did not increase BLZ2-mediated trans-activation when both type of effectors were used in the transient expression experiments (Figure 5b).

Figure 5.

 Transient expression assays of GUS activity under the control of the Hor2 and Itr1 gene promoters using HvFUS3 and BLZ2 as effectors in barley developing endosperms.
(a) Co-bombardment experiments on barley developing endosperms performed using the indicated combinations of reporter and effector plasmids at a 1:1 molar ratio. β-glucuronidase (GUS) activity in endosperms is expressed as percentage of the GUS activity for the control without effector(s). GUS expression driven by the PItr1 (wt) construct and its mutated version PItr1 (mut) is shown in the insert (white and black, respectively). In each experiment, sets of five endosperms were bombarded, and three replicate experiments were performed. Standard errors are indicated.
(b) Co-bombardment experiments on barley developing endosperms performed as in (a) but using the point mutant HvFus* (D75→N) and the deletion mutant (HvFUS3Δ7−49) as effectors.
(c) Schematic representation of the reporter and effector constructs. Reporter constructs PHor2 and PItr1 (wt) have been described previously (Díaz et al., 2002), and PItr1 (mut) is a mutated version differing at the indicated nucleotides in the RY box (see Figure 4). The position of putative cis DNA-binding sites for the FUSCA3 TF, or its mutated version, in the Itr1 promoter is indicated. Effector constructs HvFus3, HvFus3* (D75→N), HvFUS3Δ7−46 and HvBlz2 contain the indicated ORFs under the control of the CaMV 35S promoter (P35S) followed by the first intron of the maize AdhI gene (I-AdhI).

Taken together, these results indicate that HvFUS3 trans-activates the Hor2 and Itr1 gene promoters in developing barley endosperm through binding to the RY box in a combinatorial interaction with the bZIP factor BLZ2, among other TFs.

The intrinsic activation capacity of HvFUS3 protein is linked to its C-terminal region

We assayed the activation properties of HvFUS3 in the yeast one-hybrid system. In addition to HvFUS3 encoding the whole protein, a series of derived fragments containing the B3 domain, the N- and the C-terminal regions, alone or in combination with the B3 DNA-binding domain, were fused to the GAL4 binding domain (BD), and β-galactosidase activity was measured in Saccharomyces cerevisiae. As shown in Figure 6(a), cells transformed with the BD–FUS3 construct produced β-galactosidase activity of 200 Miller units, indicating that FUS3 functions as a transcriptional activator in yeast. A deletion construct, containing the C-terminal region alone, increased this activity by approximately 3.5-fold, but none of the other constructs assayed showed significant activity. This indicates the presence of an activation domain within the C-terminal region of HvFUS3, and that the inability of the mutated version FUS3Δ7−46 to trans-activate the Hor2 and Itr1 promoters (see Figure 5b) is not due to the lack of a region necessary for direct activation, but probably due to its inability to interact with other transcription factors such as BLZ2 (see below).

Figure 6.

 Yeast S. cerevisiae one- and two-hybrid system assays of HvFUS3.
(a) The indicated constructs containing the HvFus3 ORF and a series of derived fragments were used to transform S. cerevisae strain SFY526. LacZ induction was measured to detect activation domains within this protein.
(b) Effector constructs were used to transform S. cerevisae strain HF7C containing the reporter gene His3 under the control of the Gal1 promoter.
Growth of yeast cells on minimum His agar medium containing 60 mm 3-aminotriazole was assayed to test the interaction between HvFUS3, or HvFUS3Δ7−46, with BLZ2. Each colony cluster comes from a different transformation event.
(c) Protein–protein interaction of HvFUS3 and BLZ2 in the yeast two-hybrid system. The indicated effector constructs were used to transform S. cerevisae strain SFY526 containing the reporter gene LacZ under the control of the Gal1 promoter. LacZ induction was measured to test the interaction between HvFUS3 and BLZ2, and β-galactosidase activity in liquid cultures was calculated from six independent replicates in two assays. Standard errors are indicated.
HvFus3, full-length ORF; N-term, 1st to 60th amino acids of the HvFus3 ORF; B3, DNA-binding domain (61th to 171th amino acids) of the HvFus3 ORF; C-term, 172th to 285th amino acids of the HvFus3 ORF; BD, Gal4 DNA-binding domain; AD, Gal4 DNA activation domain; FUS3Δ7−46, deletion from the 7th to the 46th amino acids of the HvFus3 ORF; BLZ2, full-length HvBlz2 ORF.

HvFUS3 interacts with BLZ2 in the yeast two-hybrid system, and the HvFUS3 N-terminal region is necessary for this interaction

The co-existence of HvFus3 and HvBlz2 mRNAs in the barley endosperm, as well as the observation that trans-activation of the Hor2 and Itr1 gene promoters mediated by HvFUS3 in this tissue is increased by BLZ2, suggested that both TFs might interact in vivo. We explored this possibility in the yeast two-hybrid system using the LacZ and His3 genes as reporters. The HvFus3, HvFus3Δ7−46 and HvBlz2 ORFs were fused in-frame to yeast GAL4-BD and GaL4-AD (activation domain), and used to transform S. cerevisiae cells. As shown in Figure 6(c), expression of BD–BLZ2 was sufficient to increase reporter gene activity, consistent with the previous observation that BLZ2 possesses a transcription activation domain (Oñate et al., 1999). Co-expression of BD–BLZ2 with AD–FUS3 led to a further increase in β-galactosidase activity, demonstrating an interaction between BLZ2 and HvFUS3. This increase was not observed with the mutant HvFUS3Δ7−46. No activation of the reporter gene was observed when yeast cells were co-transformed with the two pGBT9 (BD) and pGAD424 (AD) plasmids without inserts, or with the former in combination with AD–FUS3 or AD–FUS3Δ7−46.

The interaction between FUS3 and BLZ2 was confirmed using activation of the HIS3 reporter gene and growth in His medium (Figure 6b).


FUS3 from A. thaliana is a central regulator of seed development that is involved in the establishment of dormancy, desiccation tolerance and cotyledon identity, as well as in the synthesis of SSPs and anthocyanin accumulation (Gazzarrini et al., 2004; Keith et al., 1994). Participation of FUS3 protein in these processes is coordinated within a hierarchical network involving other key regulators, namely ABI3, LEC1 and LEC2 (Kroj et al., 2003; To et al., 2006). In this study, we have isolated and functionally characterized the first FUS3-like factor from a monocot species. Our data show that barley HvFUS3 is the putative orthologue of AtFUS3, as both factors are not only phylogenetically related but can perform similar functions in distantly related systems. Like AtFUS3 in Arabidopsis seeds, HvFUS3 promotes the activation of seed-specific genes in barley. Furthermore, when introduced into a fus3 loss-of-function background in the model plant A. thaliana, it is able to restore the synthesis of SSPs and anthocyanin accumulation.

FUSCA3-mediated control of SSP genes and anthocyanin accumulation is conserved among various angiosperm species

A major goal of this investigation was to establish the possible existence of a FUS3 gene in monocot species, and to determine the degree of conserved functionality with the gene described in Arabidopsis. Following an extensive search of database sequences, we could identify the rice OsFus3, and thus isolated a barley cDNA corresponding to a gene that was the putative ortholog of AtFUS3. When the barley FUS3 cDNA was introduced into Arabidopsis fus3 plants, we observed restoration of features that are typically impaired in the mutant, such as activation of SSP gene transcription and accumulation of anthocyanins (Figure 2). Complementation of the fus3 phenotype in these respects indicates that HvFUS3 might be participating in the same regulatory network as the AtFUS3 protein, probably by interacting with other transcription factors and controlling the subsequent downstream target genes.

The capacity of HvFUS3 to activate SSP gene expression in Arabidopsis prompted us to investigate whether the FUS3 orthologous gene from barley, a monocot species, might also be involved in the transcriptional regulation of SSP and other seed-specific genes in this cereal. We addressed this question by analysis of seed gene promoters that contain conserved putative binding sites for this TF (the RY box in Figure 1) (Reidt et al., 2000) in transient expression experiments involving particle bombardment of barley developing endosperms. These experiments demonstrated that (i) HvFUS3 trans-activates the promoters of the genes Hor2 and Itr1 encoding the B-hordein storage protein and the BTI-CMe trypsin inhibitor, respectively, and (ii) the RY box is an essential element in trans-activation of these seed-specific genes. This was particularly significant in the Itr1 gene promoter, where mutation of this cis-element resulted in an 80% decrease of the GUS activity controlled by the Itr1 promoter (Figure 5).

Activation by HvFUS3 of barley SSP genes and its interchangeability with the Arabidopsis FUS3 protein is consistent with the current hypothesis of a shared origin for dicotyledonous seed globulins/albumins and monocotyledonous prolamins. Identification of HvFUS3 provides further evidence in support of the idea that evolutionary conservation affects not only protein structural domains, but it also extends to gene regulatory cis-regions and trans-acting factors (Lara et al., 2003; Vincentz et al., 1997).

HvFUS3-mediated transcriptional activation through interaction with RY boxes in target gene promoters

Arabidopsis FUS3 has been reported to activate transcription of SSP genes by binding to the RY box in their promoters (Kroj et al., 2003; Reidt et al., 2000). However, other studies show that FUS3 regulates the synthesis of SSPs through the epidermal morphogenesis regulator TTG1 or another protein functioning in a cell non-autonomous manner (Gazzarrini et al., 2004; Tsuchiya et al., 2004). Our results show that HvFUS3 binds specifically to the RY box in the promoters of the Hor2 and Itr1 genes, and activates transcription from their promoters in transient expression assays (Figures 4 and 5). Furthermore, when these assays were performed with HvFUS3*, a mutated version that is unable to bind to the RY box in EMSA experiments, the trans-activation capacity upon the Hor2 and Itr1 gene promoters was lost. We also showed in reciprocal experiments that mutations of the RY box, abolishing HvFUS3 binding to the promoter of the Itr1 gene, impaired HvFUS3-mediated trans-activation. Taken together, our data indicate that activation of these target genes in planta most likely occurs through direct interaction of FUS3 with the RY boxes present in their promoters. If a parallel scenario to Arabidopsis is assumed, the possibility cannot be ruled out that HvFUS3 might also act through an intermediate factor (TTG1 or another protein), which would direct the final interaction with RY boxes in target promoters, or even operate in a redundant manner with other B3-type TFs, such as ABI3 or LEC2. In this respect, it is worth mentioning that cereal seeds accumulate prolamins and hydrolase inhibitors in the starchy endosperm, in contrast to the SSPs of Arabidopsis, which are deposited in the cotyledons. This situation may account for potential differences in the precise mode of action of FUS3 in the two systems. In the case of barley, our data showing co-expression of HvFUS3 and prolamin genes in the starchy endosperm are in favour of a direct interaction, whereas the non-overlapping expression of FUS3 and SSPs in the embryo of Arabidopsis might require additional connections (Tsuchiya et al., 2004).

Involvement of HvFUS3 and BLZ2 in the combinatorial regulation of seed-specific gene expression

Transcriptional regulation is normally exerted by the concerted action of multiple transcription factors responding to distinct signals. Consequently, native levels of gene expression arise from the effects of various TFs acting on a particular promoter and their combinatorial interactions. The importance of such interactions (e.g. ABI3–bZIP10/25, ABI3–ABI5, VP1–TRAB, O2–PBF, GAMYB–BPBF, etc.) has already been reported in the regulation of seed-specific genes of both monocot and dicot species (Hobo et al., 1999; Lara et al., 2003; Nakamura et al., 2001).

Our transient expression assays show that co-transfection with HvFUS3 and BLZ2 increased the GUS activity above values obtained when each effector was used independently, indicating that a combination of HvFUS3 and BLZ2 is necessary for full trans-activation of the Hor2 and Itr1 genes. In addition, the TFs are able to interact in vivo in the yeast two-hybrid system. When the mutated versions HvFUS3*, lacking DNA-binding capacity (Figure 4), or HvFUS3Δ7−46, which is unable to interact with BLZ2 (Figure 6), were used in transient experiments, trans-activation of the Hor2 and Itr1 promoters did not occur (Figure 5). Moreover, the increased activation capacity expected for combination with BLZ2 was also lost in both cases, pointing to the importance of interaction of these two TFs as a requisite for their functionality.

The spatial distribution of transcripts within the barley seed is also in agreement with the regulation of target genes Hor2 and Itr1 and the importance of a combined action of HvFUS3 and BLZ2. During seed development, HvFus3 transcripts are present at significant levels in both the endosperm and the embryo (Figure 3, and data not shown). However, transcripts of Blz2 are almost undetectable in the embryo at the same developmental stage (2500-fold less than HvFus3 in the endosperm and 2000-fold less than HvFus3 in the embryo). Transcripts of the Hor2 and Itr1 genes are abundant in the endosperm (Figure 3) and nearly absent from the embryo (data not shown). The absence of BLZ2 in the embryo could explain the lack of expression of the Hor2 and Itr1 genes in embryos even at high levels of HvFUS3. However, when the promoter of one of these genes (Itr1) driving expression of the GUS reporter was introduced into tobacco plants, reporter activity also appeared in the embryo (Díaz et al., 1995). As described here for the transgenic effect of HvFUS3 in Arabidopsis, a plausible explanation is that the dicotyledonous genes corresponding to HvFUS3 and Blz2 (AtFUS3 and AtbZIP10/25 in Arabidopsis, Lara et al., 2003; Parcy et al., 1997), natively co-expressed in the embryo, are responsible for the observed pattern.

The results of our studies and those of others indicate that expression of a conserved set of SSP genes occurs in different tissues (endosperm and embryo) depending on the prevailing site of accumulation of reserves, and relies on the combined action of various regulators. Some TFs, such as FUS3, are common to both tissues and seem to have maintained this functionality within them. In this respect, the HvFUS3 regulatory activity in the cereal endosperm described here is in line with the regulation of SSP genes by FUS3 in the Arabidopsis embryo, which has also been reported for At2S3:GFP in the endosperm layer (Kroj et al., 2003). Similarly, ABI3 in dicotyledonous species and VP1 in monocots are expressed in both the embryo and the endosperm aleurone. By contrast, other TFs central to this regulation show overlapping patterns that differ from those above. For instance, the bZIP Opaque2, which is exclusively expressed in the cereal endosperm, seems to be functionally replaced by Opaque2-like bZIPs that are present in dicot embryos (Lara et al., 2003).

Interactions between B3-type and bZIP transcription factors that have important implications in seed gene regulation have been reported, both in Arabidopsis (ABI3 and AtbZIP10/25, Lara et al., 2003; ABI3 and ABI5, Nakamura et al., 2001) and in cereals (TRAB1–VP1 in rice, Hobo et al., 1999). The novel capacity of HvFUS3 to interact with bZIP factors presented in this work prompts the search for additional interactions. Comparison of such investigations between Arabidopsis and cereals is particularly attractive, as the existence of specific TFs with distinct patterns of expression opens up various possibilities for combinatorial interactions that could eventually explain the allocation of storage compounds into embryo or endosperm.

Experimental procedures

Cloning of the barley Fus3 gene, and generation of mutated constructs Fus3* and Fus3Δ7−46 and derived fragments

The ORF of the HvFus3 gene was amplified from first-strand cDNA of developing endosperm (22 d.a.p.) of barley cv. Bomi by PCR using the primers 5′-ATGGCCGCCATCAGCAGCA-3′ (primer 1) and 5′-TCACATCAGAGGCCCAGACT-3′ (primer 2) derived from the publicly available sequences BM376703 and AJ484391, respectively. The resultant amplified fragment was cloned in pGEM T-easy (Promega, http://www.promega.com/) and used for molecular characterization of the Fus3 gene and generation of its derived fragments and mutated versions. Fus3Δ7−46, corresponding to a deletion of the 7th to 46th amino acids of the ORF, was generated by PCR using primer 5′-ATGGCCGCCATCAGCAGCGGCGGTGTCCGGTGGATG-3′ and primer 2. In the mutated form of Fus3, the nucleotides encoding aspartic acid at position 75 were altered to encode asparagine, using site-directed mutagenesis with a PCR approach. The base changes were introduced in two overlapping fragments. One of them, spanning 232 bp, was amplified by PCR using primer 1 and the primer 5′-GGCTTATGTTACTGTTACGG-3′, and the other, of 646 bp, was obtained using primer 5′-CCGTAACAGTAACATAAGCC-3′ and primer 2. These fragments were annealed and extended by a PCR reaction using the two external primers 1 and 2. Other fragments derived from HvFus3 ORF were obtained by PCR using the combinations (i) primer 1 and 5′-CCCAGGCACTTGATGTGGT-3′ to amplify the N-terminal region, (ii) 5′-TTACGAGTTATTCTGCAGAAGG-3′ and 5′-ATCATCTCCCGCCTTCTTTG-3′ (primer 3) to amplify the B3 domain, (iii) primers 1 and 3 to amplify the N-terminal part plus the B3 domain, and (iv) 5′-CTAGTTGCTGCTATGCCACA-3′ and primer 2 to amplify the C-terminal domain.

Phylogenetic tree and identification of conserved motifs

The amino acid sequences of the Arabidopsis VP1/ABI3-like B3 domain genes At1g01030, At1g28300 (AtLec2), At2g30470, At2g36080, At2g46870, At3g11580, At3g24650 (AtAbi3), At3g26790 (AtFus3), At3g61970, At4g01500, At4g21550, At4g32010 and At5g06250 were obtained from http://www.Arabidopsis.org/ .The rice putative AtFus3 orthologue (Os01g51610) was obtained from http://www.tigr.org/tdb/e2k1/osa1. The amino acid sequence of the barley HvFUS3 protein was deduced using the ‘Translate’ tool at http://www.expasy.ch/. Alignment of full-length amino acid sequences through clustal w, phylogeny construction, and bootstrap tests of phylogeny by means of the neighbour-joining method and the Poisson correction model were performed using MEGA version 3.1 (Kumar et al., 2004). Conserved motifs within the amino acid sequences specified above were identified using MEME software http://meme.sdsc.edu/meme/intro.html) as described by Bailey and Elkan (1994). Default parameters were used with the following exceptions: the occurrence of a single motif was set to any number of repetitions, the maximum number of motifs to find was set to 50, and the minimum width of each motif was set to eight amino acid residues.

Expression of HvFUS3 in Arabidopsis fus3 loss-of-function and wild-type genetic backgrounds

Col-0 and fus3-3 gl1 (Luerssen et al., 1998) transgenic plants carrying a fusion of the At2S3 promoter and the coding sequence of the GFP gene have been previously described by Kroj et al. (2003). The P35S:HvFus3 and P35S:AtFus3 constructs were created by combination of the respective barley and Arabidopsis Fus3 ORFs as described by To et al. (2006). The resulting plasmids were used to transform fus3 pAt2S3:GFP plants by the Agrobacterium-mediated floral dip method (Clough and Bent, 1998). Seeds from the transformed plants were harvested and plated on medium containing Basta to select transgenic plants harbouring p35S:HvFus3 and p35S:AtFus3. Specific primers for HvFus3 and AtFus3 were used to verify the presence of the respective constructs by PCR.

Real-time quantitative RT-PCR analyses

Real-time quantitative RT-PCR studies were performed to analyse variation in mRNA accumulation for the genes HvFus3, HvBlz2, Hor2 and Itr1. Developing endosperms of barley cv. Bomi at 10, 14, 18, 22 and 26 days after pollination (d.a.p.) and immature embryos at 22 and 26 d.a.p. were isolated and collected. Total RNA was purified from frozen tissues by phenol/chloroform extraction, followed by precipitation with 3 m LiCl (Lagrimini et al., 1987) and digestion with DNase (DNase I, RNase-free; Roche Diagnostics; http://www.roche-applied-science.com). Aliquots (50 μg) of these samples were used to isolate mRNA using a QuickPrep™ Micro mRNA purification kit (GE Healthcare; http://www.gehealthcare.com), and duplicate first-strand cDNA samples were synthesized in the presence of an oligo(dT)17 primer according to the manufacturer’s instructions (first-strand cDNA synthesis kit, GE Healthcare). Real-time quantitative PCR analyses were performed for duplicate samples using an Applied Biosystems 7300 real-time PCR system (http://www.appliedbiosystems.com/) and SYBR Green detection system. Quantification was standardized to barley Actin2 mRNA levels (accession number BQ768853). The primers used for PCR amplification were as follows: barley Actin2 forward, 5′-TCCAGCTATCGTCCACAGGAA-3′; reverse, 5′-TGTCCAACAAAACCACCACT-3′; HvFus3 forward, 5′-TCTTTGTTTCCCAACCCGAA-3′; reverse, 5′-AAGGCTTGGTGACTCCGAACT-3′; HvBlz2 forward, 5′-TACATTGGTGCCTACCCCACA-3′; reverse, 5′-AATTCCGGCTCAGGTCTCTCA-3′; Hor2 forward, 5′-AGGCAATCCGTGCAATCGT-3′; reverse, 5′-TTGTCCAACTTGCTCCTGCTG-3′; Itr1 forward, 5′-TCCTCACCTCGGACATGAAGA-3′; reverse, 5′-AACGCACCCTGCCAAGTTACT-3′).

In situ hybridization

Developing barley seeds (22 d.a.p.) were collected, fixed in 4% formaldehyde in PBS, dehydrated, embedded in paraffin and sectioned to 8 μm. After de-waxing in Histoclear and rehydration, sections were treated with 0.2 m HCl, neutralized, incubated with 1 μg ml−1 proteinase K for 1 h, and dehydrated. Hybridization was performed overnight at 50°C in hybridization solution (100 μg ml−1 tRNA, 6x SSC, 3% SDS and 50% formamide), plus approximately 100 ng μl−1 of antisense or sense DIG-labelled RNA probes corresponding to a 354 bp fragment (positions 433–787 from the ATG initiation codon) of the HvFus3 gene. Two washes using 2x SSC and 50% formamide for 90 min at 50°C were then carried out. Antibody incubation and colour detection were performed according to the manufacturer’s instructions (Roche Diagnostics).

Transient expression assays in aleurone layers

The reporter constructs PHor2 and PItr1 (wt) were as described by Díaz et al. (2002). PItr1 (mut) was created using a PCR-based site-directed mutagenesis approach. The base changes in the RY box (CATGCA to ATCCTG) were introduced using two overlapping fragments. The first, spanning 263 bp, was amplified by PCR using the universal sense primer and the primer 5′-AGTGCCGCCAGGATGAAATTTTG-3′, and the second, spanning 157 bp, was obtained using primer 5′-CAAAATTTCATCCTGGCGGCACT-3′ and the GUS antisense primer (Díaz et al., 2002). These fragments were annealed and extended by a PCR reaction using the two external primers (universal sense and GUS antisense). The resulting fragment replaced the internal NotI–BamHI portion of the Itr1 promoter in the PItr1 (wt) plasmid. The effector constructs were created by cloning the whole ORFs of HvFus3 and its mutated versions Fus3Δ7−46 and Fus3* into the BamHI–EcoRI restriction sites of the pBlueScript vector (Stratagene, http://www.stratagene.com/) under the control of the CaMV 35S promoter followed by the first intron of the maize AdhI gene and the 3′nos terminator. The effector construct carrying the ORF of Blz2 is the same as that described by Oñate et al. (1999). Preparation of aleurone layers, gold particle coating and bombardment with a biolistic Helium gun device (DuPont PDS-1000, Bio-Rad, http://www.bio-rad.com/) were performed as previously described (Díaz et al., 2002). GUS expression was determined histochemically according to the method described by Jefferson (1987). GUS activity was calculated as the mean number of blue spots per aleurone in each assay and expressed as a percentage, where 100% was the value obtained using the reporter constructs without effectors. The histochemical data were directly correlated with the GUS expression quantified by chemiluminescence per milligram protein, with a correlation coefficient of 0.97 (data not shown). Chemiluminescence was assayed using a GUS Light Kit (Tropix; http://www.appliedbiosystems.com).

Yeast transformation and LacZ assays

The whole ORFs of HvFus3, its mutated versions Fus3Δ7−46 and Fus3*, and the derived fragments described above were cloned into the EcoRI–SalI restriction sites of the pGBT9 and/or pGAD424 plasmids (Clontech, http://www.clontech.com/) to generate translational fusions. The complete ORF of Blz2 fused to the Gal4 DNA-binding domain and the Gal4 activation domain in the pGBT9 and pGAD424 vectors, respectively, were as described by Oñate et al. (1999). Two haploid strains of Saccharomyces cerevisiae (Clontech) were used: SFY526, carrying a LacZ reporter gene under the control of a truncated Gal1 promoter, which contains the Gal4-responsive element, and HF7C, carrying a His3 auxotrophy gene controlled by the same promoter. Yeast transformation was performed by the polyethylene glycol method, and transformants were screened for β-galactosidase production (LacZ) and for growth in histidine-depleted agar medium (His) with 60 mm 3-aminotriazol. Quantification of β-galactosidase activity was calculated using Miller’s formula as described by Ausubel et al. (1990).

Electrophoretic mobility shift assays

The HvFUS3, HvFUS3* and AtFUS3 proteins were expressed in the E. coli BL21 CodonPlus (DE3)-RP strain (Stratagene) by cloning their ORFs into the BamHI restriction site of the pGEX-2T vector (GE Healthcare) as translational fusions to GST. Recombinant proteins were induced with 0.1 mm isopropyl-β-d-thiogalactopyranoside (IPTG) for 2 h, and bacterial extracts were obtained by sonication in 8 mm dithiothreitol (DTT) in PBS. Cells carrying the pGEX-2T vector with no insert were identically processed as negative controls. The probes containing the RY DNA-binding site derived from the Hor2, Itr1 and At2S3 promoters (Hor2-RY, Itr1-RY and At2S-3RY) and the mutated version (Itr1-ry) were produced by annealing complementary single-stranded oligonucleotides that generated 5′-protruding ends. End-labelling of the probes using [α-32P]dATP, and DNA–protein binding reactions were performed as described by Mena et al. (1998). The primers used were Hor2-RY, 5′-AAGTGAATAAGGTGAGTCATGCATACC-3′, and 3′-TTATTCCACTCAGTACGTATGGTTT-5′; Itr1-RY, 5′-TTAATCAAAATTTCCATGCAGCGGC-3′ and 3′-GTTTTAAAGGTACGTCGCCGTGAG-5′; Itr1-ry, 5′-TTAATCAAAATTTCATCCTGGCGGC-3′ and 3′-GTTTTAAAGTAGGACCGCCGTGAG -5′; At2S-3RY, 5′-ACACACTCATGCATGCATGC-3′ and 3′-GTGTACGTACGTACGTAAGTT-5′ (restriction sites underlined).


We thank Mar Gonzalez for her valuable technical assistance. Financial support from the Ministerio de Educación y Ciencia, Spain (projects BFU2006-07258 and BIO2004-00168) is gratefully acknowledged. M.A.M.-R. is the recipient of a pre-doctoral fellowship from the Comunidad de Madrid of Spain.