INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Abscisic acid (ABA) is required during seed development for the acquisition of desiccation tolerance and dormancy and plays an important role in mediating many plant responses to the environment (Busk and Pages, 1998). As ABA levels increase during the seed development program or in response to environmental stress, a number of ABA-induced genes are expressed. These include the LEA (late embryogenesis abundant) genes, which may serve to protect the developing seed from desiccation (Ried and Walker-Simmons, 1993), as well as genes such as KIN1, which encode proteins similar to antifreeze proteins (Wang and Cutler, 1995). In addition to stimulating the transcription of a suite of ABA-inducible genes, ABA also suppresses the expression of GA-induced genes encoding -amylase and Cys proteinase in aleurone layers of cereal grains (Bethke et al., 1997).

Because many physiological responses to ABA are dependent upon ABA-mediated gene expression, the signal transduction pathway involved in this process has received considerable attention (Lovegrove and Hooley, 2000; Rock, 2000). Detailed studies of the promoters of ABA-induced genes have identified an ABA response complex that is necessary and sufficient for ABA-induced transcription (Shen et al., 1996). This complex consists of the ABA response element (ABRE) (T/C)ACGTGGC together with a coupling element (CE) containing the consensus sequence CGCGTG. A number of transcription factors involved in ABA-induced gene expression have also been identified. The VP1 protein is able to activate ABA-induced genes, and requires the presence of ABREs to do so. However, VP1 does not bind directly to ABRE sequences. Members of the ABRE-binding factor (ABF) family of basic leucine zipper (bZIP) proteins do bind specifically to both ABREs and CEs (Kim et al., 1997; Choi et al., 2000) and have been shown to transactivate an ABRE-containing reporter gene in yeast (Saccharomyces cerevisiae), making them excellent candidates for the transcription factors that ultimately interact with and stimulate transcription from ABA-responsive promoters. A member of the ABF family from rice (Oryza sativa; TRAB1) interacts specifically with the VP1 protein, indicating that it probably acts to mediate VP1-dependent ABA-induced transcription (Hobo et al., 1999). A number of ABF proteins are present in Arabidopsis (Choi et al., 2000) and different members of this family may play distinct roles in the control of ABA gene expression, although the fact that many members of the family can bind to the same promoter fragment suggests that some of these roles may be at least partially redundant. The Arabidopsis ABI5 protein is also a member of the ABF family (Finkelstein and Lynch, 2000). Arabidopsis abi5 mutants have reduced seed ABA sensitivity for only a subset of ABA responses, indicating that ABI5 is responsible for mediating a particular subset of ABA responses but not others. In young seedlings, phosphorylation of ABI5 is altered in response to ABA and the presence of ABI5 is required to maintain the seedlings in a state of developmental arrest (Lopez-Molina et al., 2001). Transcriptional activation activity of two other Arabidopsis ABFs, AREB1 and AREB2, is stimulated by ABA-dependent phosphorylation (Uno et al., 2000). The protein kinases responsible for phosphorylation of the ABI5, AREB1, and AREB2 polypeptides have not yet been identified.

In addition to downstream transcription factors, proteins acting further upstream in the signal transduction pathway have also been identified. The mRNA cap-binding protein, ABH1, functions upstream of the ABA-induced rise in cytoplasmic [Ca²⁺] in Arabidopsis guard cells (Hugovieux et al., 2001). Another important component of the pathway is the ABA-induced Ser/Thr protein kinase PKABA1 (Anderberg and Walker-Simmons, 1992). PKABA1 mRNA levels increase rapidly in response to dehydration and ABA in both grains (seeds) and leaves of wheat (Triticum aestivum; Anderberg and Walker-Simmons, 1992; Holappa and Walker-Simmons, 1995). PKABA1 has been shown to act as an intermediate in the ABA antagonism of GA-induced gene expression because it was able to fully substitute for ABA in inhibiting the expression of -amylase and Cys proteinase genes in GA-treated barley (Hordeum vulgare) aleurone layers (Gómez-Cadenas et al., 1999). The effect of PKABA1 appears to be mediated through the down-regulation of GAMYB (Gómez-Cadenas et al., 2001), a transcription factor required for -amylase gene expression (Gubler et al., 1999). Further investigation is now necessary to define the precise role(s) of PKABA1 in the ABA signal transduction pathway and to identify the physiological substrate(s) of PKABA1.

To identify proteins that interact with PKABA1 in the ABA signal transduction pathway, we carried out a two-hybrid screen. We have identified TaABF, a wheat member of the ABF family that interacts with PKABA1 and contains peptide sequences that are phosphorylated by PKABA1.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Cloning of TaABF

A two-hybrid screen was used to search for cDNA clones encoding proteins that interact with PKABA1. The bait construct consisted of the complete reconstituted sequence of PKABA1 (Gómez-Cadenas et al., 1999) fused to the GAL4-binding domain (BD). This was used to screen a GAL4 activation domain (AD) fusion cDNA library made from wheat embryos imbibed for 12 h in the presence of ABA. Screening was carried out in the yeast strain PJ69-4A (James et al., 1996). This strain contains three reporter genes (ADE2, HIS3, and lacZ) that can be activated by interactions between a bait (GAL4 BD) fusion protein and a prey (GAL4 AD) fusion protein. The PKABA1 bait plasmid alone was unable to activate either the ADE2 or the HIS3 reporter genes, but was able to moderately activate the lacZ reporter gene of PJ69-4A (Fig. 1). By screening for prey cDNA plasmids activating the ADE2 and HIS3 reporter genes, we were able to identify proteins interacting with PKABA1. One of the cDNA clones identified was named TaABF(Br-2H), for reasons described below. The combination of TaABF prey plasmid and PKABA1 bait plasmid activated both the ADE2 and HIS3 reporter genes and activated the lacZ reporter gene much more strongly than the PKABA1 bait plasmid alone (Fig. 1). To confirm that TaABF bound specifically to PKABA1, it was also tested with bait plasmids encoding the peptide signaling molecule systemin (McGurl et al., 1992) and the mammalian basic helix-loop-helix transcription factor E2-2 (Henthorn et al., 1990). As shown in Figure 1, TaABF showed a positive two-hybrid interaction with PKABA1 and not with systemin or E2-2.

View larger version (53K): [in this window] [in a new window]   Figure 1.   Two-hybrid assays showing interaction between PKABA1 and TaABF. A, PJ69-4A yeast strains transformed with plasmids containing the indicated prey and bait constructs were streaked onto medium lacking the indicated nutrients. EV, Use of empty vector encoding the GAL4 BD with no fusion protein. B, Four independent transformants harboring the indicated combinations of prey and bait plasmids were assayed for -galactosidase activity using a liquid assay with O-nitrophenyl -D-galactopyranoside as the substrate. The topmost bar indicates the amount of background -galactosidase activity in untransformed PJ69-4A yeast cells. Error bars, some of which are too small to be seen, indicate SE.

The TaABF(Br-2H) cDNA clone described above contained a partial cDNA sequence (incomplete at both the 5' and 3' ends) encoding a 296-amino acid peptide fused in frame with the GAL4 AD. Additional TaABF cDNAs were obtained by screening a cDNA library from developing wheat grains. One of these cDNA clones, TaABF(CS-46), contained a 772-nucleotide region of overlap containing 100% identity with TaABF(Br-2H), and was complete at the 3' end. The cDNA sequences from TaABF(Br-2H) and TaABF(CS-46) were combined to generate the sequence of TaABFA, which is incomplete at the 5' end. The deduced amino acid sequence encoded by TaABFA is shown in Figure 2A. In addition, a full-length cDNA, TaABFB, was identified that encoded a 391-amino acid peptide that was over 95% identical (and over 98% similar) to the peptide encoded by TaABFA (Fig. 2A). The nucleotide sequence of TaABFB contains an in frame stop codon eight codons upstream of the initiator Met shown in Figure 2, indicating that it includes the entire coding region. A previously sequenced partial cDNA from wheat (BE516338.1) is completely identical to the 3' end of TaABFB. Based on comparisons with previously identified proteins in the database, the proteins encoded by these cDNAs were found to be members of the ABF family of ABRE-binding bZIP transcription factors and, thus, were designated TaABFs. Of the previously characterized members of this family, the TaABFB polypeptide is most closely related to Arabidopsis ABI5 (39% amino acid identity; Finkelstein and Lynch, 2000) and to rice TRAB1 (38% amino acid identity; Hobo et al., 1999). The amino-terminal region of the TaABFs shares three conserved sequence blocks with other members of the ABF family and it is Gly rich, as are the amino terminal regions of other ABFs. The TaABF peptides are also quite Pro rich, with a total of 24 (TaABFA) or 27 (TaABFB) Pro residues between positions 136 and 261, including a stretch of six consecutive Pro residues starting at position 246. TaABF contains a bZIP domain whose basic region contains only a single amino acid difference from those of the ABI5, TRAB1, and ABF2 (also known as AREB1) proteins (Fig. 2B). TRAB1 and ABF2 have been shown to bind both ABRE and CE promoter sequences and to activate transcription of ABA response complex-containing promoters. Other transcription factors, such as EmBP1, are able to specifically bind ABREs despite having much more divergence from TRAB1 and ABF1 in the basic domain than does TaABF.

View larger version (37K): [in this window] [in a new window]   Figure 2.   Sequence of TaABF. A, Deduced amino acid sequences from TaABF cDNA clones. The TaABFA sequence was obtained by combining the original two-hybrid clone (Br-2H, encoding amino acids 40-336) with an identical overlapping cDNA (CS-46, encoding amino acids 79-391) containing a complete 3' end. Short conserved sequence blocks present in other ABFs are underlined. The basic domain is indicated by a dashed underline and the Leu zipper region is indicated by a bold wavy underline. Amino acid residues that are identical between TaABFA and TaABFB are indicated by asterisks and similar amino acids are indicated by dots. B, Basic domain of TaABF compared with the basic (DNA-binding) regions of ABI5 (Finkelstein and Lynch, 2000), TRAB1 (Hobo et al., 1999), AtABF2, AtABF1 (Choi et al., 2000), DPBF1 and DPBF2 (Kim et al., 1997), and EmBP1 (Guiltinan et al., 1990). Nucleotide sequences are deposited in GenBank under accession numbers AF519803 (TaABFA) and AF519804 (TaABFB).

TaABF Does Not Interact with nullPKABA1 or with TaPK4

A null version of PKABA1, lacking the nucleotide-binding site that is important for protein kinase activity, has been found to be inactive in suppression of GA-induced gene expression in bombarded barley aleurone layers (Gómez-Cadenas et al., 1999). Studies of the protein kinase reaction mechanism indicate that ATP binding precedes and is required for polypeptide binding (Whitehouse et al., 1983; Cheng et al., 1998). Therefore, we constructed a nullPKABA1 bait plasmid to test whether the inactivation of PKABA1 by removal of the nucleotide-binding site would compromise its ability to bind TaABF. In a two-hybrid assay (Fig. 3A), TaABF prey plasmid was only able to activate the ADE2 reporter gene in the presence of the wild-type PKABA1 bait plasmid, but not in the presence of the nullPKABA1 bait plasmid. Similarly, the TaABF/nullPKABA prey/bait combination was unable to activate the HIS3 or the lacZ reporter genes (data not shown). These results indicate that the presence of the nucleotide-binding site of PKABA1 may be essential for binding of this protein to TaABF and provide indirect evidence that TaABF may be a physiological substrate for PKABA1.

View larger version (41K): [in this window] [in a new window]   Figure 3.   Two-hybrid assays of TaABF with PKABA1, nullPKABA1, and TaPK4. A, PJ69-4A yeast strains transformed with plasmids containing the indicated prey and constructs were streaked onto medium lacking the indicated nutrients. The TaABF prey plasmid used was the TaABF(Br-2H) obtained in the two-hybrid screen. EV, Use of empty vector encoding the GAL4 BD with no fusion protein. B, Complete deduced amino acid sequences of TaPK4 and PKABA1. The first 10 amino acids of the PKABA1 clone reconstituted from the genomic clone (see "Materials and Methods") are italicized. The nucleotide-binding site of PKABA1 that is absent in nullPKABA1 is indicated by a double underline. The C-terminal acidic stretches in TaPK4 and PKABA1 are indicated by a single underline. Amino acid residues that are identical between TaPK4 and PKABA1 are indicated by asterisks. The nucleotide sequence of TaPK4 is deposited in GenBank under accession number AF519805.

In addition to PKABA1, wheat contains at least two additional members of the SnRK2 subfamily (also known as PKABA1 subfamily) of protein kinases (Halford and Hardie, 1998). One of these, TaPK3, has 97% amino acid identity to PKABA1 (Holappa and Walker-Simmons, 1997). We have recently identified an additional member of this family, TaPK4, which has only 55% amino acid identity with PKABA1 (Fig. 3B). TaPK4 is more closely related to two Arabidopsis protein kinases, ASK1 and ASK2, whose mRNA is up-regulated by light (Park et al., 1993) and, thus, it is more likely to carry out a function that is distinct from that of PKABA1. As shown in Figure 3A, TaABF does not bind to TaPK4 in a two-hybrid assay, suggesting that PKABA1, but not TaPK4, may be connected to TaABF in the ABA signal transduction pathway.

Expression of TaABF

Figure 4 shows that TaABF mRNA was present in whole mature grains. TaABF mRNA was not induced, however, in either leaves or roots of stressed seedlings. Cutting and drying of leaves, a treatment that has been shown previously to induce high levels of ABA (Holappa and Walker-Simmons, 1995), also failed to induce TaABF gene expression. In contrast to TaABF, both PKABA1 and TaPK4 mRNAs were induced by the cutting and drying treatment. Although both PKABA1 and TaPK4 were stress induced, their patterns of expression were clearly distinct because PKABA1 mRNA was more abundant in mature grains than in cut leaves, whereas TaPK4 mRNA was much more abundant in cut leaves than in mature grains.

View larger version (109K): [in this window] [in a new window]   Figure 4.   Transcript levels of TaABF, PKABA1, and TaPK4 in stressed wheat plants. Seven-day-old wheat seedlings were maintained under control conditions (C) of 22°C and 100% relative humidity or were subjected to a cold treatment of 2°C (2°), application of 250 mM NaCl to the roots (S), application of 25 µM ABA to the roots (A), or removal to a drier chamber maintained at 85% relative humidity (D). After 24 h of the stress treatments, leaves and roots were separately collected for analysis. In a separate experiment, the top 4 cm of leaves was removed from 7-d-old seedlings and placed at 85% relative humidity for the indicated number of hours. RNA was also obtained from whole mature, after-ripened grains. Ten micrograms (20 µg for the TaPK4 blot) of total RNA was electrophoresed, blotted, and hybridized to the cDNA probes indicated on the right side of the figure. The TaABF probe used was the TaABF(Br-2H) partial cDNA, which would be expected to hybridize with both TaABFA and TaABFB transcripts. Total RNA was detected by ethidium bromide fluorescence.

The finding that TaABF mRNA was present in mature grains along with PKABA1 mRNA led us to investigate the pattern of expression of these two mRNAs in developing and germinating grains. As shown in Figure 5A, both PKABA1 and TaABF mRNAs accumulated at the end of the grain development program. These transcripts began to reach detectable levels by around 30 DAP and then continued increasing in abundance until the grains reached maturity at 45 DAP. These results confirm earlier work (Anderberg and Walker-Simmons, 1992) showing that PKABA1 mRNA was present in maturing grains. Brevor grains are dormant at maturity, and unless subjected to a period of after ripening (storage under warm dry conditions), they will not germinate during a 3-d period of imbibition. We investigated the levels of PKABA1 and TaPK4 mRNAs in both after-ripened and dormant grains during imbibition. We found that the levels of the two transcripts decreased during imbibition of after-ripened grains, which completed germination in about 20 h (Fig. 5B). In contrast to this pattern, TaABF transcripts transiently increased from 6 to 24 h of imbibition in the dormant grains. To determine whether PKABA1 and TaABF mRNAs were present in the same grain tissues, we assayed for the presence of these transcripts in total RNA purified from dissected endosperms (including the aleurone layer) and embryos. As shown in Figure 5C, PKABA1 mRNA was present at relatively equal levels in both the embryo and endosperm, whereas TaABF mRNA was more abundant in the endosperm than in the embryo.

View larger version (47K): [in this window] [in a new window]   Figure 5.   Transcript levels of TaABF and PKABA1 in wheat grains. A, Grains were collected from greenhouse-grown wheat plants at 5 to 45 d after pollination (DAP). Grains collected at 45 DAP were fully mature. B, Grains were stored after harvest for 1 year at room temperature to obtain after-ripened (AR) grains. Grains maintained at 20°C for 1 year after harvest were used for the dormant grains. Grains were placed on moist filter paper and allowed to imbibe for 0 to 48 h. C, After-ripened grains were allowed to imbibe for 3 h before dissection of embryo (emb) and endosperm (end) tissue or collection of whole grains. Ten micrograms of total RNA was electrophoresed, blotted, and hybridized to the cDNA probes indicated on the right side of the figure. The TaABF probe used was the TaABF(Br-2H) partial cDNA. Total RNA was detected by ethidium bromide fluorescence.

Phosphorylation of TaABF1 Sequences by PKABA1

If TaABF is a physiological substrate of PKABA1, then PKABA1 should be able to phosphorylate specific peptide motifs of TaABF in vitro. Because it was not technically feasible to obtain purified PKABA1 that retained protein kinase activity, crude protein extracts from cell lines containing PKABA1 were used. Extract from a Drosophila melanogaster cell line transformed with a PKABA1 construct was prepared and was determined by immunoblotting (data not shown) to contain a low level of PKABA1 protein. Six synthetic peptides (see Fig. 6A), representing different motifs from the TaABFA peptide sequence, were tested for their ability to be phosphorylated. These peptides represent motifs that are distinctive of ABF proteins and contain Ser or Thr residues. The SAMS peptide (HMRSAMSGLHLVKRR; Davies et al., 1989), a substrate for AMP-activated protein kinase, was also tested.

View larger version (22K): [in this window] [in a new window]   Figure 6.   In vitro phosphorylation of TaABFA-derived peptides by PKABA1. A, Sequence of synthetic peptides used as substrates in in vitro phosphorylation assays. Residues representing possible phosphorylation sites for PKABA1 (S,T) are in bold. Basic residues added at the end of synthetic peptides for technical reasons, but not present in the TaABFA sequence, are underlined. The MNM peptide represents amino acid residues 60 to 73 of TaABFA as numbered in Figure 2. The other peptides represent residues 102 through 123 (VW), 163 through 178 (GEM), 254 through 274 (SR), 291 through 310 (SCER), and 311 through 326 (BD). B, Phosphorylation of synthetic peptides by control extract and FLAG::PKABA1 extract. A control cell line (C) and a cell line transformed with a gene encoding a FLAG-PKABA1 fusion protein (P) were used for the assays. Crude protein extract was prepared from the cell line and used in in vitro phosphorylation assays in the presence of -32P ATP and 50 µg of the indicated peptide for 15 min. After removal of molecules >10,000 D from the reaction products, phosphorylated peptides were bound to P81 paper and the amount of phosphorylation measured by scintillation counting. Error bars indicate SE. D, Phosphorylation of BD peptide by control and PKABA1 extracts. Assays were carried out as in B except that the amount of peptide included was varied.

These peptides were tested utilizing both a cell line transformed with a PKABA1 construct and a control D. melanogaster cell line lacking any transgene. Phosphorylation of the SAMS peptide was observed; however, the amount of SAMS phosphorylation by the PKABA1 cell line was no greater than by the control cell line, indicating that a component(s) of the control extract was able to phosphorylate the SAMS peptide, but that PKABA1 did not do so. For peptides VW, GEM, and BD, the amount of phosphorylation by the PKABA1-containing extract was clearly increased over that carried out by the control extract, indicating that PKABA1 phosphorylated these peptides.

To further investigate the phosphorylation of peptides by the cell extracts and to clarify the role of PKABA1 versus the components present in control cell extracts, we tested both the PKABA1-containing extract and the control extract in the presence of varying amounts of the BD peptide. At lower amounts of peptide, relatively little phosphorylation was carried out by the control cell extract (Fig. 6C). As the amount of peptide was increased, a gradual increase in the amount of phosphorylation was observed, suggesting that a component of the control extract with a low specificity for the BD peptide could phosphorylate BD in the presence of high peptide concentrations. In contrast, the PKABA1 extract was able to maximally phosphorylate BD at much lower peptide concentrations, indicating that a component of this extract (which must be PKABA1) was able to phosphorylate the BD peptide with high affinity. When 5 µg of peptide was used in the assay, the phosphorylation of BD by the PKABA1 extract was 8-fold greater than the control extract. This is strong evidence that the PKABA1 protein in this extract provided a clearly increased level of phosphorylation over that of the control.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

We have determined that TaABF, a member of the ABF family of bZIP factors, interacts with the ABA-responsive protein kinase PKABA1. TaABF contains several conserved motifs that are typical of the ABF family, suggesting that TaABF is an ABRE-binding transcription factor. TaABF contains the three conserved N-terminal sequence blocks that are distinctive for ABF proteins, as well as a bZIP domain that is very highly conserved with other members of the ABF family such as ABI5 and TRAB1. The fact that a considerable range of diversity exists in the basic (DNA-binding) domains of proteins that have been shown to bind ABREs (e.g. compare TRAB1 with EmBP1 in Fig. 2B) suggests very strongly that TaABF, which contains only one amino acid difference from TRAB1 or ABF2, is itself an ABRE-binding transcription factor. We have identified two slightly different TaABF cDNAs (TaABFA and TaABFB) that encode very similar, but not identical, peptides. Members of the ABF family (Choi et al., 2000) are generally only conserved within the three N-terminal sequence blocks and in the bZIP domain. Thus, the very close similarity between TaABFA and TaABFB indicates that they may be encoded by orthologous genes from different genomes within the allohexaploid wheat genome.

Previously identified members of the ABF family of bZIP transcription factors have been found to be regulated in a variety of ways. The Arabidopsis ABF2 (also known as AREB1) and ABF4 (also known as AREB2) genes are up-regulated in response to ABA and a variety of stresses, at the mRNA level (Choi et al., 2000; Uno et al., 2000). Activation or modification of ABFs has also been observed at the protein level. Binding of the soybean (Glycine max) bZIP factor SGBF-1 to ABREs and SGBF-1-mediated gene expression are both enhanced by binding of the zinc finger protein SCOF-1 to SGBF-1 (Kim et al., 2001). In addition to their regulation at the mRNA level, both ABF2 and ABF4 have been observed to be phosphorylated by an unidentified protein kinase in response to ABA (Uno et al., 2000). This phosphorylation may serve to regulate the ability of ABF2 and ABF4 to bind ABREs or to activate transcription of ABA-responsive genes. Although the data are less clear for ABI5, it appears that the phosphorylation state of this protein is also altered in response to ABA (Lopez-Molina et al., 2001).

The ability of TaABF to interact with PKABA1 suggests that it might be an important physiological substrate for PKABA1 and that phosphorylation of TaABF may regulate its activity. Mechanistic studies with the model Ser/Thr protein kinase cAMP-dependent protein kinase have indicated that its substrates are bound in a preferred order. The kinase first binds ATP at the nucleotide-binding site. This binding induces a conformational change in the kinase that allows subsequent binding of the protein or peptide substrate (Whitehouse et al., 1983; Cheng et al., 1998). If binding of ATP at the PKABA1 nucleotide binding site must precede binding of the protein substrate, then it would be expected that elimination of the nucleotide-binding site would prevent PKABA1 from binding its physiological protein substrate. The fact that only active PKABA1, and not null PKABA1 (which lacks the nucleotide binding site), was able to bind TaABF strongly suggests that PKABA1 binds TaABF as a phosphorylation substrate rather than in some other less specific manner. The specificity of the interaction between TaABF and PKABA1 is further supported by the observation that TaABF does not interact with TaPK4, another member of the SnRK2 subfamily that is present in wheat. The fact that the original two-hybrid cDNA clone, TaABF(Br-2H), encodes only amino acids 40 to 336 of the 391-amino acid TaABF protein indicates that this central region is sufficient for binding by PKABA1.

RNA-blot analysis indicates that TaABF expression is grain specific and that its mRNA accumulates during late grain development and disappears during imbibition of germinating seeds. Under all the tested conditions in which TaABF mRNA was present, PKABA1 mRNA was also present. However, TaABF mRNA was only present during a subset of the conditions and tissues in which PKABA1 mRNA was present. It is possible that another, as yet unidentified, member of the wheat ABF family, or some other type of factor, may serve as the physiological substrate for PKABA1 in stressed leaf and root tissues. The fact that the expression patterns of TaABF and TaPK4 do not overlap provides further evidence that TaABF is not a physiological substrate for TaPK4. TaPK4 is a newly discovered protein kinase that has 55% amino acid identity with PKABA1. TaPK4 mRNA is up-regulated in leaves in a manner very similar to that of PKABA1. TaPK4, however, is expressed at very low levels in wheat grains, whereas PKABA1 mRNA accumulates during late seed development to levels at least as high as those seen in stressed leaves. The role of TaPK4 in ABA signal transduction, if any, has yet to be determined. If it is a part of this pathway, it apparently functions in vegetative tissues and belongs to a distinct branch of the pathway that does not involve TaABF. We have classified TaPK4 as a probable protein kinase based on sequence similarity to previously identified kinases. TaPK4 contains most of the conserved subdomains (I-XII; Hanks and Quinn, 1991) and "invariant" residues that are typically found in the catalytic domain of protein kinases. However, subdomain I is only partially conserved and a conserved Lys present in subdomain II of most protein kinases is absent from TaPK4. There is as yet no biochemical evidence for kinase activity of TaPK4 or any knowledge of its biological substrate, and further experimentation will be necessary to determine if TaPK4 does have protein kinase activity.

If TaABF is a physiological substrate of PKABA1 that participates in ABA signal transduction, then PKABA1 should be able to phosphorylate TaABF in vitro. Because Ser/Thr protein kinases generally recognize phosphoacceptor sites based on local features, short peptides representing the phosphoacceptor sites in the substrate protein are regularly used to assay protein kinase activity (Ruzzene and Pinna, 1999). We used six synthetic peptides, representing potential phosphorylation sites on TaABF, as substrates in kinases assays. PKABA1 is apparently quite labile during protein purification procedures because attempts to isolate PKABA1 from transformed D. melanogaster cell lines resulted in a loss of protein kinase activity. We were able to show, however, that cell lines producing PKABA1 exhibited significantly higher phosphorylation of the VW, GEM, and BD peptides than did control cell lines.

The VW and GEM peptides represent two of the three conserved sequence blocks found in all members of the ABF family. A recent study by Uno et al. (2000) indicated that recombinant Arabidopsis ABF2 (also known as AREB1) and ABF4 (also known as AREB2) protein fragments representing these conserved sequence blocks were phosphorylated by an unidentified 42-kD protein kinase. A recombinant fragment representing the bZIP domain of ABF4 was not observed to be phosphorylated under the same conditions. The BD peptide derived from TaABF represents sequences within the basic DNA BD and in vitro phosphorylation of this peptide by PKABA1 suggest that the basic domain of TaABF may be phosphorylated in vivo by PKABA1. For three of the synthetic peptides used (MNM, GEM, and SR), there are slight differences in the peptide sequences from the corresponding regions of TaABFA versus TaABFB. For example, the TLGEMTLE motif (from TaABFA) that is phosphorylated by PKABA1 is present as TLGELTLE in TaABFB. Whether this conservative amino acid substitution has any effect on PKABA1-mediated phosphorylation of TaABF will have to be resolved by further experiments.

Although the role of PKABA1 in ABA signal transduction is still not fully understood, significant strides have been made in recent years. It is now known that PKABA1 is involved in the ABA-mediated repression of GA-stimulated -amylase gene expression in barley aleurone layers (Gómez-Cadenas et al., 1999). ABA results in the increased production of PKABA1, which then acts to suppress the transcription of GAMYB, a transcription factor that stimulates -amylase transcription (Gómez-Cadenas et al., 2001). The mechanism by which PKABA1 suppresses GAMYB transcription is not currently known. It is also not well established whether PKABA1 has a role in the stimulation of ABA-induced gene expression.

It is likely that the expression of specific genes is important in the maintenance of seed dormancy during imbibition (Johnson et al., 1995; Li and Foley, 1997). The increase in TaABF gene expression during imbibition of dormant wheat grains suggests that TaABF may play an important role in the suppression of germination-related events. PKABA1 is also present in imbibing dormant seeds, and has been shown to be involved in the ABA-mediated suppression of GA-induced gene expression during imbibition and germination. Therefore, it is possible that in response to ABA, PKABA1-mediated phosphorylation of TaABF could lead to suppression of germination-associated gene expression and germination itself. The fact that ABI5 is responsible for growth arrest in Arabidopsis seedlings suggests that TaABF may play a similar role in imbibing wheat grains. We hypothesize that a high level of TaABF in the grain suppresses germination and GA-responsive gene expression until environmental conditions are favorable for seedling establishment.

TaABF is likely to be a key component of the ABA signal transduction pathway that serves as a link between PKABA1 and downstream effects on gene expression. It may also play a critical role in maintaining dormancy in wheat grains, a condition that is extremely important for food quality and economic value of harvested wheat crops. Future studies of the effects of TaABF overexpression and underexpression in cells responding to ABA, using models such as the barley aleurone transient expression system (Gómez-Cadenas et al., 2001), should help to further clarify the role of TaABF in ABA signal transduction.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Two-Hybrid Screening and Assays

Two-hybrid screening (Chien et al., 1991) was carried out using the Matchmaker two-hybrid system (CLONTECH, Palo Alto, CA) according to the manufacturer's instructions. The bait plasmid containing the PKABA1 open reading frame fused to the GAL4 BD utilized a reconstituted PKABA1 coding sequence consisting of the PKABA1 cDNA (Anderberg and Walker-Simmons, 1992) with additional 5' sequence information (encoding the first 10 amino acids) from a genomic clone (Holappa and Walker-Simmons, 1997). An NdeI/BamHI fragment containing the entire reconstituted PKABA1 coding region was ligated into NdeI/BamHI-digested pAS2-1 (CLONTECH) to generate the bait plasmid. pAS2-1/null PKABA1 has the nucleotide-binding site (GSGNFG, amino acids 11-16) deleted (Gómez-Cadenas et al., 1999). pAS2-1/TaPK4 contains the coding region of TaPK4 cloned into the BamHI site of pAS2-1. The GAL4 AD cDNA fusion library was constructed in pGAD10 by CLONTECH using poly(A⁺) mRNA from wheat (Triticum aestivum cv Brevor) embryos that had been imbibed for 12 h (22°C) in the presence of 25 µM ABA in 5 mM MES, pH 5.7.

Yeast (Saccharomyces cerevisiae) strain PJ69-4A (James et al., 1996) containing the HIS3, ADE2, and lacZ reporter genes was used for all experiments. Yeast transformations were carried out as described by (Gietz and Schiestl, 1995) and cells were plated onto SC medium (James et al., 1996) lacking Trp, Leu, and adenine. After 4 d, positive colonies were picked and plated onto SC medium lacking Trp, Leu, and His and containing 1 mM 3-aminotriazole. Plasmids were isolated from colonies autotrophic for both adenine and His and were used to electrophorate Escherichia coli HB101 cells. Prey plasmids were isolated from HB101 colonies growing on minimal medium lacking Leu. cDNA inserts were then sequenced using an automated sequencer (ABI, Foster City, CA). Purified prey plasmids were then retransformed into yeast with and without bait plasmids to confirm the specificity of interactions. Liquid -galactosidase assays were carried out according to CLONTECH Matchmaker protocols using O-nitrophenyl -D-galactopyranoside as the substrate. Sequence analysis was carried using the BLAST program (Altshul et al., 1990) and Gene Inspector (Textco Inc., West Lebanon, NH). Additional TaABF cDNAs were obtained by screening a cDNA library obtained from developing wheat grains (25-45 DAP), using the TaABF(Br-2H) cDNA as a probe. The TaPK4 cDNA was obtained from a cDNA library obtained from wheat embryos that had been imbibed for 12 h (22°C) in the presence of 25 µM ABA in 5 mM MES, pH 5.7.

RNA Analysis

Wheat seedlings were grown at 22°C and 100% relative humidity as previously described (Holappa and Walker-Simmons, 1995). For the stress treatments, whole 7-d-old seedlings were placed in a chamber at 2°C (cold), at 35% relative humidity (dry), or were left in the 22°C chamber and watered with 250 mM NaCl or 25 µM ABA. After 24 h of the stress treatments, leaf and root tissue were collected separately and immediately frozen in liquid nitrogen. For the cutting/drying time course, the top 4 cm of leaves was collected from 7-d-old seedlings and were placed in a chamber maintained at 22°C and 85% relative humidity (Holappa and Walker-Simmons, 1995). After the appropriate dehydration period, the leaves were immediately frozen in liquid nitrogen.

Total RNA was extracted from the samples using the RNeasy Midi Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions for purification of plant RNA. The only exception was that grain tissues (1 g) were initially ground in a mortar and pestle and homogenized for 1 min in a mixture of 6 mL of 200 mM Tris (pH 8), 100 mM LiCl, 5 mM EDTA, and 1% (w/v) SDS and 6 mL of phenol:chloroform. The homogenate was incubated at 50°C for 20 min and centrifuged at 3,000g for 10 min. The top aqueous layer was then extracted with an equal volume of chloroform and a crude RNA pellet was precipitated by addition of 0.1 volumes of 0.3 M sodium acetate and 1 volume of isopropanol. After centrifugation for 15 min at 5,000g, the pellets were dissolved in 500 µL of water and then added to 5 mL of Qiagen buffer RPE to continue with the RNeasy procedure. RNA blotting and probe preparation was carried out as previously described (Johnson et al., 1995), using TaABF(Br-2H) as a probe. Blots were hybridized in Ultrahyb (Ambion, Austin, TX) at 50°C according to the manufacturer's instructions and washed in 0.1× SSC and 0.1% (w/v) SDS at 65°C. Blots were then exposed to a phosphor screen and analyzed using a phosphor imager.

Expression of PKABA1 Constructs in Drosophila melanogaster Schneider 2 Cells

PKABA1 protein was produced using Schneider 2 cells from the D. melanogaster Expression System version B (Invitrogen, Carlsbad, CA). PKABA1 and FLAG::PKABA1 expression vectors were prepared by insertion of the appropriate construct into the pMT/V5-His (C) vector. Transformed D. melanogaster cultures were grown in medium containing 10% (w/v) fetal calf serum and 300 µg mL¹ hygromycin-B and maintained at a cell density of approximately 6 × 10⁶ cells mL¹. Production of PKABA1 protein was induced in the presence of 500 µM copper sulfate for 30 h. The induced cells were harvested by gentle centrifugation, washed with phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 1.8 mM KH₂PO₄, pH 7.4), resuspended in 500 µL of cold extraction buffer (50 mM Tris, pH 7.4; 50 mM KCl; 50 mM NaF; 4 mM EDTA; 2 µg mL¹ antipain, 1 µg mL¹ leupeptin; 1 µg mL¹ pepstatin; and 100 µg mL¹ phenylmethylsulfonyl fluoride), and lysed by sonication at 0°C. The cell debris was removed by centrifugation and total protein concentrations of the supernatants were measured using the Bradford assay (Bradford, 1976).

In Vitro Peptide Phosphorylation

Synthetic peptides were synthesized at the Washington State University Laboratory for Biotechnology and Bioanalysis, using fluorenylmethoxylcarbonyl chemistry on a 431A peptide synthesizer (PE-Applied Biosystems, Foster City, CA). After each coupling step, the uncoupled amine on the resin was acetylated to prevent the formation of any peptides with internal deletions. Enzymatic activity of the cell extracts containing PKABA1 was measured using an in vitro peptide phosphorylation assay (McMichael et al., 1995). The level of kinase activity on endogenous D. melanogaster proteins was minimized through removal of phosphorylated proteins via column separation and peptide purification on cation exchange discs. In vitro peptide phosphorylation reactions (30 µL) consisted of 5 µg of total soluble protein fraction, varying amounts (0-50 µg) of peptide (dissolved in water), label buffer (40 mM Tris, pH 8.0; 10 mM MgSO₄; and 5 mM MgCl₂) and 3 µCi of -³²P-ATP, and were performed at 25°C, with rocking for 15 min. The reaction was stopped by heating at 100°C for 3 min. The volume was diluted to a total volume of 100 µL and added to a Centricon Y10 (10-kD cutoff) spin column (Amicon, Beverly, MA) for separation of the labeled peptide from any labeled endogenous proteins. The filtrate was collected containing the labeled peptide and 50-µL aliquots were applied to phosphocellulose P81 (Whatman, Hilsboro, OR) discs. Unincorporated -³²P-ATP was removed by washing with three rinses of 75 mM phosphoric acid followed by 95% (w/v) ethanol. After drying, the amount of radioactivity on each disc was determined by scintillation counting.