Identification and Characterization of a Novel Cap-binding Protein from Arabidopsis thaliana *

Cap-binding proteins specifically bind to the 7-methyl guanosine (m7G) functional group at the 5′ end of eukaryotic mRNAs. A novel Arabidopsis thalianaprotein has been identified that has sequence similarity to cap-binding proteins but is clearly a different form of the protein. The most obvious primary sequence difference is the substitution of two of the eight conserved tryptophan residues with other aromatic amino acids in the novel protein. Analogous forms of this novel protein appear to be present in other higher eukaryotes but not in yeast. Analysis of the native and recombinant forms of the novel protein by retention on m7GTP-Sepharose indicate that it is a functional cap-binding protein. Measurements of the dissociation constant for this protein indicate that it binds m7GTP 5–20-fold tighter than eukaryotic initiation factor (eIF)(iso)4E. The novel protein also supports the initiation of translation of capped mRNA in vitro. Biochemical analysis and yeast two-hybrid data indicate that it interacts with eIF(iso)4G to form a complex. Based on these observations, this protein appears to be able to function as a cap-binding protein and is given the designation of novel cap-binding protein (nCBP).

bind the m 7 G functional group found at the 5Ј end of most eukaryotic cellular mRNAs (for recent reviews, see Refs. 1 and 2). The binding of eIF4E is thought to be the first step in the assembly of several initiation factors on the mRNA prior to the binding of the 40 S ribosome (1)(2)(3). One of the structural features of all eIF4E proteins is the presence of eight tryptophan residues in the same relative positions. Recently, the cocrystal structures of the mouse eIF4E (4) and yeast eIF4E (5) with m 7 GDP were solved. Tryptophan residues 3, 5, and 8 were found to be involved in binding the m 7 G functional group (4,5). eIF4E is usually associated with a larger subunit (eIF4G) in a complex named eIF4F (1,6). eIF4A may also be present in the eIF4E⅐eIF4G complex depending upon the method of purification (7,8). Higher plants possess a unique second form of eIF4F designated eIF(iso)4F (9). eIF(iso)4F contains distinct forms of the cap-binding protein (eIF(iso)4E) and the larger subunit (eIF(iso)4G) (10). The two isoenzyme forms of plant eIF4F have the same activities in vitro (9); however, the eIF(iso)4E prefers hypermethylated caps and mRNAs with less secondary structure (11,12). Recently, a second form of mammalian eIF4G was reported (13). However, this form of eIF4G was not reported to have a distinct form of eIF4E associated with it and does not appear to be the functional equivalent of eIF(iso)4G.
The mammalian eIF4E is known to be phosphorylated at Ser-209, and the phosphorylation state appears to correlate with activity of the protein (reviewed in Refs. 14 and 15). Certain stimuli, including insulin and several growth factors, induce phosphorylation of eIF4E (14). Overexpression of the mammalian eIF4E results in cell transformation, suggesting that eIF4E levels play a critical role in normal growth and/or development (16,17). The recent discovery of proteins that specifically bind mammalian eIF4E and sequester it from interacting with eIF4G have linked the insulin signaling pathway directly to translation (reviewed in Refs. 14 and 18). These eIF4E binding proteins (4E-BP) are a major target for phosphorylation following insulin or growth factor treatment, and the phosphorylated form of 4E-BP no longer binds eIF4E (18). Thus the availability and phosphorylation state of eIF4E are crucial regulatory mechanisms for translational control in mammals. It is not known if similar types of regulation by phosphorylation or sequestration occur in plants.
In this report, we have identified a novel cap-binding protein (nCBP) from Arabidopsis thaliana that is distinct in both amino acid sequence and m 7 GTP binding properties from eIF4E or eIF(iso)4E and appears to be present only in higher eukaryotes.

EXPERIMENTAL PROCEDURES
Materials-The cDNA (19) encoding the expressed sequence tag (EST) for the nCBP (77G6T7) was obtained from the Arabidopsis Biological Resource Center (ABRC, Ohio State University). DNA sequencing of the cDNA (both strands) was performed at the DNA sequencing facility of the Institute for Cellular and Molecular Biology (University of Texas at Austin). Expression vector pET15b and Escherichia coli DE3(HMS174) were obtained from Novagen (Madison, WI). Restriction enzymes and DNA modifying enzymes were obtained from Life Technologies Inc. IPTG and X-gal were from Ambion (Austin, TX) or Research Products International (Mt. Prospect, IL). The A. thaliana suspension culture was generously provided by A. N. S. Reddy (Colorado State University). Preparation of the Arabidopsis extracts and chromatography on m 7 GTP-Sepharose (Pharmacia Biotech Inc.) will be described elsewhere but were similar to the preparation and fractionation of wheat germ extracts (20). Wheat eIF(iso)4E and eIF(iso)4G were prepared as described previously (21). Protein determinations were by the method of Bradford (22). SDS-PAGE (12.5 or 20% acrylamide) and Western blots were carried out as described previously (23). The silver stain kit was from Novex (San Diego, CA). Antibodies to recombinant Purification of Novel Cap-binding Protein-The coding region for Arabidopsis nCBP was amplified with the appropriate primers containing XhoI and BamHI sites. The amplified DNA was cloned into pET15b restricted with XhoI and BamHI and in frame with the 6-histidine tag. The plasmid construct was confirmed by DNA sequencing. The plasmid containing the nCBP coding region was transformed into E. coli DE3(HMS174). Expression of the recombinant protein was carried out as described for eIF(iso)4E, except the incubation temperature of the culture after induction was at 25°C (21). The cell pellet from a 1.5-liter culture was resuspended in 15 ml of buffer C (20 mM Hepes-KOH, pH 7.6, 10% glycerol) containing 0.4 M KCl. The cells were sonicated and centrifuged for 30 min at 30,000 rpm in a Ti-60 rotor, and the supernatant was applied to a Ni-NTA column (Qiagen, Valencia, CA) equilibrated with the same buffer. The column was washed with buffer C containing 0.1 M KCl and 40 mM imidazole and then eluted with buffer C containing 0.1 mM KCl and 400 mM imidazole. The fractions containing the highest amount of protein were pooled and dialyzed against buffer B (20 mM Hepes-KOH, pH 7.6, 10% glycerol, 1 mM DTT, 0.1 mM EDTA) containing 0.1 M KCl. nCBP prepared by Ni-NTA chromatography was used to raise rabbit antibodies.
Alternatively for purification on m 7 GTP-Sepharose, the cell pellet was resuspended in 10 ml of 20 mM Hepes-KOH, pH 7.6, 20% glycerol, 1 mM DTT, 0.1 mM EDTA containing 0.4 M KCl. The cells were sonicated and treated as described above, and the supernatant was applied to a 2-ml m 7 GTP-Sepharose column (capacity ϳ1.5 mg). The column was washed with 20 mM Hepes-KOH, pH 7.6, 20% glycerol, 1 mM DTT, 0.1 mM EDTA, 0.1 M KCl. The nCBP was eluted with the same buffer solution containing 400 M m 7 GTP. The fractions containing the highest amount of protein were pooled and dialyzed against 20 mM Hepes-KOH, pH 7.6, 20% glycerol, 1 mM DTT, 0.1 mM EDTA, 0.1 M KCl.
In Vitro Polymerization Assay-The activity of nCBP was measured in a wheat germ translation system (100 l) dependent upon the addition of eIF(iso)4G and eIF(iso)4E (21). The indicated amounts of either nCBP or eIF(iso)4E were added to the reaction mixture containing 20 pmol of eIF(iso)4G. The amount of [ 14 C]leucine incorporated was determined as described previously (20). Capped ␤-globin mRNA was prepared as described (24).
Fluorescence Measurements-Fluorescence measurements were carried out at 23°C in 20 mM HEPES, pH 7.6, 1 mM MgCl 2 . The concentration of protein was 0.25 M for both nCBP and eIF(iso)4F. The data were collected and analyzed as described previously (25). Excitation was at 280 nm and emission monitored at 330 nm. Background fluorescence was subtracted.
Affinity Chromatography of nCBP and eIF(iso)4F on m 7 GTP-Sepharose-A mixture containing 130 g of recombinant nCBP and 130 g of recombinant wheat eIF(iso)4G was applied to a 0.5 ml column of m 7 GTP-Sepharose equilibrated in buffer B containing 0.1 M KCl, and the flow-through fractions (400 l) were collected. The matrix was washed with an additional 5 ml of buffer. The retained protein was eluted with buffer B containing 0.3 M KCl and 400 M m 7 GTP, and fractions (400 l) were collected. The flow-through and eluate fractions were analyzed by SDS-PAGE.
Yeast Two-hybrid Analysis-The coding regions of nCBP and Arabidopsis eIF(iso)4E were placed into the binding domain plasmid, pGBT9 (26), after amplification of the coding region with the appropriate primers. The constructs were confirmed by DNA sequencing. The wheat eIF(iso)4E in pGBT9/N and wheat eIF(iso)4G in pGAD424 were prepared as described previously (26). The ␤-galactosidase color assay with X-gal was performed according to the protocol of the manufacturer.

RESULTS AND DISCUSSION
Analysis of the EST data base for A. thaliana cap-binding proteins, eIF4E and eIF(iso)4E, revealed a sequence that appeared to be a novel form of eIF4E. This novel cap-binding protein, termed nCBP, differed significantly in sequence from other known eIF4E or eIF(iso)4E sequences. One of the obvious differences was the substitution of two of the conserved tryptophan residues (Trp-1 and -3) with other aromatic residues (Fig. 1A). Further inspection of the EST data base revealed many mouse and human ESTs that were more similar to the Arabidopsis nCBP than to mammalian eIF4E. The number of independent mouse and human nCBP ESTs that appear in the EST data base suggest that the mammalian nCBP mRNAs are highly expressed. This prediction was confirmed by Northern blot analysis of RNA from various mouse tissues that showed high levels of nCBP mRNA expression in all tissues. 2 Temeles et al. (27) reported several mouse cDNA sequences for transcripts that were highly expressed in preimplantation embryonic mouse tissue, but the protein products were not all identified. One of these unidentified cDNAs (GenBank™ DataBank accession number U01137) appears to encode the mouse nCBP protein (Fig. 1B). Sequence comparison of the nCBP from Arabidopsis and mouse shows that these two proteins are more similar to each other (note Trp-1 and -3 are altered also in mouse nCBP) than to their cognate normal cap-binding proteins (compare Fig. 1, panels A and B). Two of the non-conservative amino acid changes in the Arabidopsis nCBP are also present in the mouse nCBP, e.g. L 136 3 E and T 171 3 N (numbering for Arabidopsis nCBP). These two amino acids are absolutely conserved among eIF4E from plants, yeast, Drosophila, Xenopus, and mammals, suggesting that the mouse and Arabidopsis nCBP are more related to each other than to their cognate eIF4E.
Three of the conserved tryptophans (Trp-3, -5, and -8 in Fig.  1A) were recently shown to be involved in the binding of m 7 GDP by solving the structures of mouse eIF4E and yeast eIF4E with bound m 7 GDP (4,5). Mutagenesis studies with yeast (28) and human (29) eIF4E indicated that mutation of the conserved Trp residues (with the exception of Trp 4 in yeast) either eliminated or reduced the ability of eIF4E to bind m 7 G affinity columns. It was, therefore, of great interest to determine if the nCBP, which is lacking Trp-1 and -3, is able to bind m 7 G functional groups.
An E. coli extract containing the expressed recombinant nCBP was applied to a m 7 GTP-Sepharose column to test its ability to interact with m 7 G functional groups. The m 7 GTP-Sepharose column (2.0 ml) has an estimated capacity of 1.5 mg. We find that fewer nonspecific proteins are retained on the column when the capacity of the column is slightly exceeded. Consequently, only 70% of the soluble recombinant nCBP remaining after ultracentrifugation was retained on the column as judged by densitometry (Fig. 2, lanes 1 and 2). The nCBP was eluted with m 7 GTP (Fig. 2, lanes 3-8) and approximately 1.2 mg of protein was recovered. It was observed that the nCBP appeared to elute less well from the m 7 GTP-Sepharose column than either eIF(iso)4E or eIF4E. Increasing the concentration of m 7 GTP from 100 M to 400 M appeared to enhance the elution and the recovery. This observation suggested that the nCBP may have a more avid interaction with m 7 G functional groups.
Initial measurements using fluorescence spectroscopy indicate at least a 5-20-fold difference in the binding affinity of nCBP and eIF(iso)4E for m 7 GTP (Fig. 3). The high affinity of m 7 GTP for nCBP precludes accurate determination of the binding constant by fluorescence due to limiting signal at low protein concentrations. Nonetheless, the evident difference in binding affinities provides an explanation for the difficulty in elution of nCBP from m 7 GTP-Sepharose. Further analysis with other cap analogs and capped oligonucleotides will be essential to elucidate the basis for the differences in observed binding constants.
To obtain an estimate of the relative amount of nCBP present in Arabidopsis, an extract was prepared from Arabidopsis suspension cultures and applied to a m 7 GTP-Sepharose column. Fig. 4A shows a silver-stained gel of the m 7 GTP eluate (lane 1) and purified recombinant nCBP (lanes 2 and 3). The Arabidopsis eIF4E and eIF(iso)4E in the eluate were identified by Western blot analysis using mouse antiserum raised to recombinant wheat eIF(iso)4E (Fig. 4B, lane 1)   spectrometry identification of peptides (data not shown). The amount of nCBP present in the m 7 GTP eluate (Fig. 4A, lane 1) appears to be at least 10-fold lower than that of either eIF4E or eIF(iso)4E, suggesting that this protein is present in very low amounts in Arabidopsis. nCBP in the Arabidopsis m 7 GTP eluate was identified with antiserum to recombinant nCBP (Fig.  4B, lane 2). Antibody raised to recombinant wheat eIF(iso)4E cross-reacts with both Arabidopsis eIF4E and eIF(iso)4E (Fig.  4B, lane 1), indicating that these proteins are closely related. However, note that there is no cross-reaction between the nCBP and other Arabidopsis cap-binding proteins with either the antiserum to nCBP or the antiserum to wheat eIF(iso)4F (Fig. 4B, lanes 1 and 2). The lack of cross-reaction between nCBP and eIF4E or eIF(iso)4E, and the amino acid sequence data suggest that nCBP is a more distantly related form of a cap-binding protein.
To further test the interaction of the nCBP with m 7 G functional groups, the ability of nCBP to translate a capped mRNA was determined. As shown in Fig. 5A, the nCBP is able to substitute for wheat eIF(iso)4E, in the presence of wheat eIF(iso)4G. The nCBP is able to support initiation of translation at about 30% of the level of eIF(iso)4E. This result implies that nCBP interacts with a capped mRNA but also forms a functional complex with eIF(iso)4G. We have also obtained data that show nCBP is also able to interact with eIF4G (data not shown). It is not clear at this time whether the lower protein synthesis activity is an innate characteristic of nCBP itself or is due to some other unknown reason (e.g. not having the correct binding partner).
The inhibition of translation initiation by m 7 GTP in the presence of eIF(iso)4E or nCBP was also determined. As shown in Fig. 5B, the nCBP appears to be much less sensitive to the presence of m 7 GTP in the translation assay than eIF(iso)4E. At least 5-6-fold more m 7 GTP was required to obtain 50% inhibition. The order of addition of the m 7 GTP did not appear to have any affect on the level of inhibition. 3 This result implies that the interaction of capped ␤-globin mRNA is much stronger with the nCBP than with eIF(iso)4E and therefore requires higher amounts of m 7 GTP to dissociate the complex. Additional fluorescence binding experiments with capped RNAs will be necessary to determine the dissociation constants relative to that of m 7 GTP.
The ability of nCBP to support polypeptide synthesis in the 3 K. A. Ruud, S. R. Lax, and K. S. Browning, unpublished data. presence of eIF(iso)4G implies the ability to form a complex. If the two proteins form a stable complex, the complex will be retained on m 7 GTP-Sepharose. We have shown previously (21) that eIF(iso)4G is not retained on m 7 GTP-Sepharose except in the presence of eIF(iso)4E. Recombinant wheat eIF(iso)4G and recombinant nCBP were incubated together and applied to a m 7 GTP-Sepharose column. Elution of the column with m 7 GTP shows that both the nCBP and eIF(iso)4G were bound (Fig. 6). These results confirm that the recombinant nCBP interacts with eIF(iso)4G to form a functional complex. Further evidence of the interaction of the nCBP with eIF(iso)4G was obtained in the yeast two-hybrid system. We have previously shown that wheat eIF(iso)4G and eIF(iso)4E interact strongly in the yeast two-hybrid system (26). The nCBP and Arabidopsis eIF(iso)4E were placed into a binding domain vector (pGBT9/N) and cotransformed into yeast with wheat eIF(iso)4G in an activation domain vector (pGAD424). Specific interaction with the nCBP and wheat eIF(iso)4G was observed (Table I) as well as interaction of the control wheat or Arabidopsis eIF(iso)4E with wheat eIF(iso)4G. No interaction of nCBP with eIF4A or eIF4B was observed in this system (data not shown). These results further confirm the potential for nCBP to specifically interact with eIF(iso)4G or a similar protein.
We have identified a new and novel protein that functions as a cap-binding protein and can interact with eIF(iso)4G to form a complex that supports the protein synthesis initiation of a capped mRNA. This protein appears to be ubiquitous in higher eukaryotes. ESTs for the nCBP are present for maize, Brassica napus, mouse, and human; also, a Caenorhabditis elegans predicted gene product appears to be very similar to the nCBP. Interestingly, there does not appear to be a homolog of this protein in yeast. Careful inspection of the yeast genome did not reveal a hidden gene that is similar to the nCBP. Therefore nCBP may be necessary for a function specific to multicellular organisms (i.e. differentiation). The mouse nCBP cDNA was among a group of cDNAs found in a screen for mRNAs that were highly expressed in preimplantation embryos (27), a tissue undergoing rapid differentiation.
The precise biological role of the nCBP is not obvious. The higher binding affinity of nCBP for m 7 GTP would imply a discriminatory role for the nCBP protein during initiation of translation. However, the lower activity of nCBP in translation suggests that a possible function may be to sequester mRNAs and either slow down or prevent their translation. We have been unable to demonstrate any competition in vitro of the nCBP with eIF4E or eIF(iso)4E (data not shown). However, we have preliminary evidence that the translation of some mRNAs is not supported by the nCBP (data not shown), suggesting that discrimination and/or sequestering may be the functional role of nCBP. We are currently determining the expression levels of nCBP mRNA and protein in various Arabidopsis tissues as well as identifying additional proteins that interact with nCBP. This additional data may provide more information about the function of this novel protein in the initiation of translation or regulation of gene expression in higher eukaryotes.