The Phosphorylation State of the Wheat Translation Initiation Factors eIF4B, eIF4A, and eIF2 Is Differentially Regulated during Seed Development and Germination*

The translation initiation factors (eIF) 4B and eIF2 are phosphoproteins whose phosphorylation state differs between mature seed and leaves. We examined the isoforms of eIF4B and the α and β subunits of eIF2 during the development and germination of wheat seed to determine whether the differences in their phosphorylation state are because of tissue-specific regulation or occur concomitant with changes in protein synthetic activity during development. eIF2α underwent phosphorylation through several intermediate isoforms that correlated with the increase and subsequent reduction in protein synthetic activity characteristic of seed development. eIF2β and eIF4B, present as highly phosphorylated isoforms during early seed development, underwent dephosphorylation during late development. eIF4B was rapidly phosphorylated within 20 h of germination, whereas eIF2α did not undergo dephosphorylation until 48–60 h of growth. A third factor, eIF4A, was predominantly nonphosphorylated throughout most of seed development and germination. These observations suggest that the phosphorylation state of eIF2α, eIF2β, and eIF4B is developmentally regulated in a way that correlates with the changes in protein synthetic activity but that some differences were also observed.

The translation initiation factors (eIF) 4B and eIF2 are phosphoproteins whose phosphorylation state differs between mature seed and leaves. We examined the isoforms of eIF4B and the ␣ and ␤ subunits of eIF2 during the development and germination of wheat seed to determine whether the differences in their phosphorylation state are because of tissue-specific regulation or occur concomitant with changes in protein synthetic activity during development. eIF2␣ underwent phosphorylation through several intermediate isoforms that correlated with the increase and subsequent reduction in protein synthetic activity characteristic of seed development. eIF2␤ and eIF4B, present as highly phosphorylated isoforms during early seed development, underwent dephosphorylation during late development. eIF4B was rapidly phosphorylated within 20 h of germination, whereas eIF2␣ did not undergo dephosphorylation until 48 -60 h of growth. A third factor, eIF4A, was predominantly nonphosphorylated throughout most of seed development and germination. These observations suggest that the phosphorylation state of eIF2␣, eIF2␤, and eIF4B is developmentally regulated in a way that correlates with the changes in protein synthetic activity but that some differences were also observed.
Protein synthesis undergoes dramatic changes during plant growth, particularly during the development of the seed and its subsequent germination. From the perspective of protein synthetic activity, the development of wheat seed can be divided into three stages: early development (up to 10 -12 days after flowering (DAF) 1 ), mid-development (approximately 12-30 DAF), and late development (30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45). During early development, the endosperm and embryo are surrounded by the nucellus, a maternal tissue that provides nutrients to the rapidly growing endosperm (1,2). The nucellus is the predominant tissue during early seed development but undergoes a progressive programmed cell death 2 to provide room for the expanding endosperm tissue. Following 12 DAF, the nucellus is insignificant and the mass of the seed is dominated by the endosperm and embryo. The highest level of protein synthetic activity in the seed occurs between 12 and 30 DAF when the bulk of the seed storage protein is synthesized in the endosperm and stored to provide the nutrients required for germination (3)(4)(5). This developmental stage is followed by a rapid drop in protein synthesis when the endosperm undergoes its own programmed cell death and the embryo prepares for quiescence, a process that continues up to the mature dry stage of the seed by 45 DAF. Despite the reduced protein synthetic activity after 30 DAF, late embryogenesis abundant (lea) proteins are synthesized, which are thought to be required for the embryo to survive dessication during late seed development (reviewed in Ref. 6). At maturity, however, the seed is virtually metabolically quiescent, and translational activity ceases despite the continued presence of the translational machinery in the embryo. Protein synthesis resumes quickly following the initiation of germination; formation of polysomes and protein synthesis are detected shortly following the onset of germination (7)(8)(9)(10).
Such extreme differences in protein synthetic activity during development in plants presumably requires considerable control over the activity of the translational machinery itself. Translational regulation in yeast and animal cells often occurs through changes in the phosphorylation state of the eukaryotic initiation factors (eIFs) whose function is to aid 40-and 60-S ribosomal subunit binding to the mRNA, resulting in the formation of the 80-S ribosome at the correct start codon. Our previous studies demonstrated that the phosphorylation state of two initiation factors, eIF2 and eIF4B, differed substantially in seed versus leaves of wheat (11), observations suggesting that the activity of these factors may be under developmental regulation.
Wheat eIF4B, a single polypeptide (59 kDa), is thought to assist eIF4A and eIF4F in the ATP-dependent unwinding of secondary structure that may be present within a 5Ј-leader of an mRNA (12,13). eIF4B is hypophosphorylated in mature wheat embryos, but is present predominantly as hyperphosphorylated isoforms in leaves (11). The posttranslational modification of eIF4B is similar in both plant and animal cells in that the isoforms are divided into an acidic cluster of four to six isoforms that are phosphorylated forms of a basic cluster of four isoforms (11,14). The acidic cluster is substantially reduced in mature wheat embryos, but predominates in leaves, and undergoes rapid dephosphorylation in leaves following a heat shock (11). The alteration in wheat eIF4B phosphorylation correlates with the repression of translation that occurs following a thermal stress (15). Further evidence suggesting a relationship between the phosphorylation state of eIF4B and translational activity comes from studies in animal cells: dephosphorylation of mammalian eIF4B occurred following heat shock (16), serum depletion (17), or mitosis (18). Dephos-phorylation of mammalian eIF4B correlated with the reduction in translation following these treatments, whereas the phosphorylation of eIF4B that occurs following insulin treatment correlated with an increase in translation (19). Moreover, the addition of phosphorylated mammalian eIF4B was partially able to restore translation in an in vitro translation lysate prepared from heat-shocked HeLa cells in which the eIF4B had undergone dephosphorylation (16,20).
eIF2 is a three-subunit complex in plants as in other eukaryotes and is responsible for binding the initiator Met-tRNA to the 40-S ribosomal subunit (21)(22)(23). The three subunits of wheat eIF2 are ␣ (42 kDa), ␤ (38 kDa), and ␥ (50 kDa) (24). Phosphorylation of eIF2␣ occurs following amino acid starvation in yeast (reviewed in Ref. 25) or viral infection (reviewed in Ref. 26), heme deprivation, and heat shock in animal cells (27). Its phosphorylation inhibits the eIF2␤-directed exchange of GTP for GDP, preventing eIF2 from participating in more than one round of translation initiation (reviewed in Refs. 28 -30). In wheat, eIF2␣ is also subject to phosphorylation: the factor is present in a hyperphosphorylated state in embryos of mature seed but is present in a hypophosphorylated state in leaves (11). However, in contrast to mammalian eIF2, heat shock, whether of short or long duration, had little detectable impact on the phosphorylation state of the ␣ subunit in wheat leaves (11). Mammalian eIF2␤ is present as a single major form under normal growth conditions (14), but following either a heat shock or serum starvation, both a new acidic and basic species was observed (14,17). In contrast, approximately four eIF2␤ isoforms were observed in wheat leaves, the distribution of which did not change significantly following a heat shock (11). eIF4A (47 kDa) is an ATP-dependent RNA helicase and an RNA-dependent ATPase in plants, animals, and yeast (31,32). It can be found associated with eIF4G or eIFiso4G or as an independent factor (33,34). Only a single, dephosphorylated form of mammalian or yeast eIF4A has been observed; however, in plants and Drosophila, a phosphorylated isoform has been detected (34,35). Although dephosphorylated eIF4A normally predominates in leaves and mature seed, a maximum of 50% of the total available eIF4A can undergo phosphorylation following hypoxia or heat shock (11,34).
Our previous observations showing that the phosphorylation state of several initiation factors is different in seed versus leaves of wheat (11) suggested that this was because of either tissue-specific differences or to differences in translational activity between mature seed (translationally quiescent) and young leaves (translationally active). This prompted us to examine whether the posttranslational modification of these factors may be subject to developmental control that would correlate with the regulation of protein synthetic activity known to occur during development. We found that the phosphorylation state of eIF4B, eIF2␣, and eIF2␤ (and to a lesser extent, eIF4A) is developmentally regulated during seed development and germination but that the regulation of each differs temporally. The differences in the regulation of the phosphorylation state of these initiation factors may affect message selection or determine translational efficiency during plant development.

EXPERIMENTAL PROCEDURES
Antibody Preparation-eIF4A and eIF2 (31) were purified from wheat germ (commercially prepared wheat embryos) as described. The purification of recombinant eIF4B and eIF4A will be described elsewhere. 3 Polyclonal antibodies to eIF2 and recombinant eIF4B and eIF4A were produced in rabbits and affinity-purified as described (33,36). Antibodies raised against purified PABP were prepared as described previously (37).
Plant Extract Preparation and Two-dimensional Gel Electrophore-sis/Western Blot Analysis-For the seed development studies, wheat plants were grown in a greenhouse to minimize any potential environmental effects and were allowed to self-pollinate. Whole seeds were collected at various stages of seed development and frozen in liquid nitrogen. For the germination studies, wheat seed or seedling tissues were dissected at various stages of germination and frozen. Total soluble protein extracts were prepared by grinding tissue in a mortar with liquid nitrogen and then in aqueous buffer (50 mM HEPES, pH 7.5, 120 mM KOAc, 5 mM MgOAc, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 g/ml pepstatin, 1 mM EDTA, and 0.5 mM okadaic acid). The cell debris was pelleted by centrifugation and the protein concentration determined (38). Depending on the factor examined, 5-100 g of protein was loaded on IEF tube gels (3.6% acrylamide, 9 M urea, 2% ampholytes, 2% Nonidet P-40) and run at 400 V for 4.5 h followed by 0.5 h at 500 V. The protein was then resolved in the second dimension using standard SDS-PAGE and the protein transferred to 0.22 m nitrocellulose membrane by electroblotting. Following transfer, the nitrocellulose membranes were blocked in 5% milk, 0.01% thimerosal in TPBS (0.1% Tween 20, 13.7 mM NaCl, 0.27 mM KCl, 1 mM Na 2 HPO 4 , and 0.14 mM KH 2 PO 4 ) followed by incubation with primary antibodies diluted 1:2000 for eIF4A and 1:1000 for eIF4B, eIF2␣, and eIF2␤ in TPBS with 1% milk for 1.5 h. The blots were then washed twice with TPBS and incubated with goat anti-rabbit horseradish peroxidase-conjugated antibodies (Southern Biotechnology Associates, Inc.) diluted to 1:10,000 for 1 h. The blots were washed twice with TPBS, and the signal detected was typically between 1 and 15 min using chemiluminescence (Amersham Pharmacia Biotech). The range of ampholytes used is indicated in the legend to each figure. The pH range of the IEF gel following isoelectric focusing was determined from the measurements of 5 mm sections of a control gel soaked in 1 ml of 15 mM NaCl.

eIF4B Undergoes Dephosphorylation Late in Seed Development and Is Phosphorylated Early during Germination-To
examine whether the phosphorylation state of eIF4B changes during seed development, wheat seed were collected at developmental stages following pollination. Soluble protein extracts prepared from whole seeds were resolved on 2D gels and analyzed using antibodies raised against recombinant eIF4B following Western blotting. The early development of wheat seed is characterized by a small endosperm and embryo and a large nucellus. In the early developmental time points (4 -12 DAF), intermediate acidic isoforms of eIF4B were present with a second set of lower molecular weight isoforms also present ( Fig.  1). This apparent degradation of eIF4B correlated with the programmed cell death of nucellar tissue that occurs at this stage in the development of the seed (T. Young and D. Gallie, manuscript in preparation). This observation is supported by the presence of only the full-length acidic isoforms of eIF4B at the next developmental stage (16 DAF) by which point the nucellus had virtually disappeared. Also by 16 DAF, the distribution of eIF4B isoforms had shifted to highly phosphorylated (acidic) species and remained unchanged through 30 DAF which corresponds to the period in which the bulk of protein synthesis takes place (3)(4)(5). However, at 30 DAF, the basic cluster of eIF4B isoforms appeared for the first time, representing a redistribution of the isoforms away from phosphorylated to dephosphorylated species, a process that continued through late seed development. By 45 DAF, the phosphorylated species had disappeared altogether, leaving only the basic cluster. The eIF4B isoforms were identical in the endosperm and embryo at this stage (data not shown), despite the fact that the embryo is viable whereas the endosperm has undergone programmed cell death (T. Young and D. Gallie, manuscript in preparation), data suggesting that dephosphorylation of eIF4B during development is not a consequence of cell death.
We next examined whether the distribution of eIF4B isoforms was regulated during germination. Imbibing seed were collected at time points and the embryo dissected. The endosperm was not examined as this is a dead tissue in germi-nating seed. Upon emergence of the shoot and roots at later stages of germination, these tissues were also collected and the samples were analyzed as described above. eIF4B was present as predominantly hypophosphorylated isoforms in seed before imbibition and up to 10 h of germination (Fig. 2). By 20 h, a shift to more phosphorylated isoforms was observed that was completed by 48 h of germination when the shoot had first emerged. eIF4B present in roots of 48 h germinated seed was also present as highly phosphorylated isoforms (Fig. 2). The observed shift of eIF4B from hypophosphorylated to hyperphosphorylated species during germination correlates with the resumption of protein synthesis in wheat (7-10).
eIF4A Is Phosphorylated at the Nucellus-containing Stage of the Young Seed but Not during Subsequent Seed Development or during Germination-The same extracts were used to determine whether the phosphorylation state of eIF4A changes during seed development and germination. Dephosphorylated eIF4A largely predominates in wheat leaves or maize roots under normal conditions but undergoes phosphorylation to a maximum 1:1 ratio following a severe heat shock or hypoxic treatment (11,34). During early seed development when the nucellus predominates and undergoes programmed cell death (4 -12 DAF), the two isoforms of eIF4A were present in approximately equimolar amounts (Fig. 3). When this maternal tissue had disappeared by 16 DAF, the dephosphorylated form of eIF4A once again predominated and the distribution of the isoforms remained little altered during the remaining development of the seed. This observation suggests that the phosphorylated form of eIF4A is present in the nucellus and may be targeted for inactivation or degradation during programmed cell death. During germination, only the dephosphorylated isoform was observed in the embryo (data not shown) and up through 2 day-old shoots (Fig. 4). However, the presence of both eIF4A isoforms was observed in 2 day-old roots (Fig. 4).
eIF2␣ Undergoes Progressive Phosphorylation during Seed Development and Is Dephosphorylated during Germination-In the embryo of mature seed, the ␣ subunit of eIF2 is present in its most phosphorylated state (pI of 5.4 -5.6) whereas in leaves it is completely dephosphorylated (pI of 6.2-6.4) (11). This large difference in pI is in striking contrast to that observed in animal cells in which the ␣ subunit is subject to a single phosphorylation event at serine-51 (reviewed in Ref. 39) resulting in a small shift in the pI of the subunit from 5.4 to 5.6 (14,40). This suggests that the ␣ subunit of eIF2 in plants may be multiply phosphorylated. The number of ␤ subunit isoforms also appears to differ between plants and FIG. 1. eIF4B undergoes dephosphorylation during late seed development in wheat. 60 -100 g of soluble protein extract from developing wheat seed taken at the indicated days after flowering (DAF) was resolved in the first dimension using IEF with 75%, pH 5-8, and 25%, pH 3-10, ampholytes and on an 8% SDS-PAGE gel for the second dimension. eIF4B degradation products can be seen in the first three time points. The acidic and basic clusters of the eIF4B isoforms are each indicated with a bracket above the top panel.
FIG. 2. eIF4B undergoes progressive phosphorylation during wheat germination. 10 -30 g of soluble protein extract from wheat germ, embryos excised from germinating seed (0, 1, 5, 10, and 20 h after imbibition), or from leaves and roots (48 h) was resolved in the first dimension using IEF with 75%, pH 5-8, and 25%, pH 3-10, ampholytes and on an 8% SDS-PAGE gel for the second dimension. Recombinant eIF4B, purified following its expression in Escherichia coli, was included in the analysis. The eIF4B acidic and basic clusters are each indicated with brackets.
FIG. 3. Two-dimensional gel electrophoresis/Western analysis of the change in phosphorylation state of eIF4A during seed development. 5 g of soluble protein extract from developing wheat seed taken at the indicated days after flowering (DAF) was resolved in the first dimension using IEF with 50%, pH 5-8, and 50%, pH 2.5-5, ampholytes and on a 10% SDS-PAGE gel for the second dimension. The eIF4A phosphorylated isoform is indicated with a downward pointing arrow.
animals. Whereas a single major mammalian eIF2␤ species is observed under normal growth conditions (14), at least four eIF2␤ isoforms were observed in wheat leaves (11). We examined, therefore, whether multiple isoforms of the ␣ and ␤ subunits were present in wheat seed and whether their distribution changed during seed development. As in leaves, approximately four major species were observed for the ␤ subunit during early seed development (Fig. 5), although the distribution in seed included more acidic isoforms not detected in leaves. The two most basic species present at 4 -12 DAF (Fig. 5) correspond to the middle two, and most dominant, species in leaves (11). With further development of the seed, a progressive redistribution toward the basic isoforms was observed such that by 35 DAF, only a single basic isoform predominated. This eIF2␤ basic isoform serves as a convenient reference point (see arrows, Fig. 5) for the changes in the eIF2␣ subunit isoforms described below. In contrast, small amounts of some of the phosphorylated eIF2␤ isoforms observed during mid-development remain present in mature embryos (see 0 h, Fig. 7 and (11)), suggesting tissue-specific (endosperm versus embryo) regulation of its phosphorylation state during late development.
At the earliest stages of seed development examined (4 -12 DAF), multiple isoforms of eIF2␣ were detected (Fig. 6) but mostly were basic to intermediate acidic species and those at 12 DAF corresponded to those dephosphorylated species previously observed in young wheat leaves (11). By 16 DAF, a shift to slightly more acidic species was observed (Fig. 6). The most basic of the two ␣ isoforms observed at 16 DAF corresponded to the most acidic of the isoforms present at 12 DAF. The distribution of eIF2␣ isoforms remained unchanged up to 20 DAF. By 25 DAF, the isoforms that had been present at 16 -20 DAF had disappeared and a new set of 3-4 highly acidic isoforms (pI range from 5.4 to 5.9) appeared, the most acidic of which (pI 5.4) corresponded to the fully phosphorylated ␣ subunit isoform that was observed in mature wheat embryos (11). At 30 DAF, there was an increasing shift of the ␣ subunit to the most acidic isoform and by 40 DAF, only the most phosphorylated isoform remained. At 40 DAF, a second, smaller molecular weight form of the ␣ subunit was observed in addition to the full-length form. This was also observed with eIF2 purified from mature wheat embryos (11) and may be because of degradation of the ␣ subunit during late seed development. At 45 DAF, the ␣ subunit was no longer detectable despite the fact that the ␤ subunit was still present, data suggesting that the ␣ subunit is degraded during late endosperm development. In contrast, eIF2␣ remains present in mature embryos (see 0 h, Fig. 7 and  (11)). These data demonstrate that the eIF2␣ subunit is present in a hypophosphorylated state during early seed development but undergoes a progressive shift to a highly phosphoryl-ated state that begins at 16 DAF, is completed by 35 DAF, and is followed by degradation specifically in the endosperm at 45 DAF but not completely in mature embryos.
Upon imbibition, the ␣ subunit was present solely in its highly phosphorylated form in the embryo (Fig. 7). No change in eIF2 phosphorylation in the embryo was observed between 10 and 20 h of germination, the point at which eIF4B was observed to undergo phosphorylation (see Fig. 3). Between 48 and 60 h of growth (see Figs. 7 and 8, respectively), the basic isoforms of the ␣ subunit are detectable in the emerging shoot. FIG. 4. eIF4A is differentially phosphorylated in shoot and root tissue of wheat seedlings. 5 g of soluble protein extract from leaves and roots (48 h) was resolved in the first dimension using IEF with 50%, pH 5-8, and 50%, pH 2.5-5, ampholytes and on a 10% SDS-PAGE gel for the second dimension. The eIF4A phosphorylated isoform is indicated with a downward pointing arrow.
FIG. 5. eIF2␤ undergoes conversion from multiple phosphorylated isoforms to a single hypophosphorylated species during seed development. 30 g of soluble protein extract from developing wheat seed taken at the indicated days after flowering (DAF) was resolved in the first dimension using IEF with 75%, pH 5-8, and 25%, pH 3-10, ampholytes and on a 12% SDS-PAGE gel for the second dimension. Affinity-purified, anti-␤ subunit antibodies were used to probe for the ␤ subunit. The ␣ and ␤ subunits of eIF2 are indicated to the left of the gels. The large arrow in each panel points to the most abundant ␤ subunit isoform.
FIG. 6. eIF2␣ undergoes progressive phosphorylation during seed development. Affinity-purified, anti-␣ subunit antibodies were used to reprobe the membrane from Fig. 5 so that ␤ subunits would also be evident to serve as points of reference. The ␣ and ␤ subunits of eIF2 are indicated to the left of the gels. The ␣ subunit isoforms are indicated by arrows.
However, substantial levels of the most phosphorylated isoforms remained without the appearance of the intermediate species, suggesting that the two extreme forms of the ␣ subunit temporally coexist during this stage of germination. This differs substantially from the observations made during seed de-velopment where the conversion of the ␣ subunit from the hypophosphorylated to hyperphosphorylated state occurred en masse and the two extreme forms of the ␣ subunit were temporally separate. Only by 4 days of growth had the ␣ subunit been completely converted to the basic isoforms (Fig. 8). Similar results were observed in the roots of 4 day-old seedlings where the ␣ subunit was present solely in its dephosphorylated form. Germination in the light promoted the conversion of eIF2␣ to its dephosphorylated state as dark-germinated seedlings contained nearly equimolar amounts of phosphorylated and dephosphorylated ␣ subunit (Fig. 9). These data show that eIF2␣ subunit undergoes a shift from being present in a highly phosphorylated state during early germination to a hypophosphorylated state upon emergence of the shoot that occurs over a period of several days.
In contrast, phosphorylation of eIF2␣ occurs earlier during seed development than does the dephosphorylation of eIF4B, and eIF2␣ undergoes dephosphorylation much later during germination than does eIF4B. Interestingly, conversion of eIF2␣ to a hyperphosphorylated state during seed development occurred in a concerted manner, suggesting either that its phosphorylation occurs in a distributive fashion or that phosphorylation of each site is under temporal control during development. During germination, however, dephosphorylation of eIF2␣ occurred without the significant appearance of the intermediate isoforms, even though both the hyperphosphorylated and hypophosphorylated isoforms were present simultaneously. This suggests that either eIF2␣ dephosphorylation occurs in a processive manner or that its phosphorylation is irreversible, and dephosphorylated eIF2␣ is generated only through new synthesis.
The presence of multiple eIF2␣ phosphorylation sites in plants appears to differ from the single site present at serine-51 in mammalian eIF2␣ (41,42). As in mammalian cells, eIF2␣ activity in yeast is also regulated through phosphorylation at Ser-51, however, three additional sites close to the C terminus are constitutively phosphorylated in vitro and in vivo by casein FIG. 7. eIF2␣ begins conversion to its dephosphorylated state only during late germination. 100 g of soluble protein extract from embryos excised from germinating wheat seed (0, 1, 10, and 20 h after imbibition) or from leaves and roots (48 h) was resolved in the first dimension using IEF with 75%, pH 5-8, and 25%, pH 3-10, ampholytes and on a 12% SDS-PAGE gel for the second dimension. Affinity-purified, anti-␤ subunit antibodies were used to probe for the ␤ subunit (left column of panels), and the membranes were reprobed with anti-␣ subunit antibodies (right column of panels). The ␣ subunit isoforms are indicated by arrows. FIG. 8. eIF2␣ undergoes complete dephosphorylation only during late germination. 100 g of soluble protein extract from 1-day-old embryos, shoots of 2-, 2.5-, 3-, and 4-day-old germinated wheat, or roots of 4-day-old seedlings was resolved in the first dimension using IEF with 75%, pH 5-8, and 25%, pH 3-10, ampholytes and on a 12% SDS-PAGE gel for the second dimension. The membranes were then probed for the ␣ and ␤ subunits. The ␣ subunit isoforms are indicated by arrows. FIG. 9. Dephosphorylation of eIF2␣ during germination is promoted by growth in light. 100 g of soluble protein extract from 3-day-old shoots germinated in light or in complete darkness was resolved in the first dimension using IEF with 75%, pH 5-8, and 25%, pH 3-10, ampholytes and on a 12% SDS-PAGE gel for the second dimension. The membranes were then probed for the ␣ and ␤ subunits. The ␣ subunit isoforms are indicated by arrows.
kinase II (43) and remain so following heat shock, nitrogen starvation, or growth in poor carbon sources (44). Mutations at these sites do not reduce growth, but can exhibit synthetic growth defects in combination with mutations affecting the activity of eIF2B (the factor required for GTP/GDP exchange in eIF2) or GCN2 (the kinase responsible for phosphorylating eIF2␣), suggesting that their phosphorylation is necessary for optimal activity of the factor (43). Although human eIF2␣ does not contain consensus sites for and is not phosphorylated by casein kinase II, Artemia eIF2␣ contains four such sites and is phosphorylated by the enzyme in vitro (43,45), suggesting considerable variation in the number of phosphorylation sites of eIF2␣ from various species.
None of the phosphorylation sites in plant eIF2␣ can be considered as constitutively phosphorylated since the most hypophosphorylated species dominates in leaves and during early seed development. How phosphorylation at each site of wheat eIF2␣ affects its activity remains to be determined. However, as changes in eIF2␣ phosphorylation correlate with the substantial changes in protein synthetic activity during seed development, they do suggest a possible role for phosphorylation as a means to regulate eIF2␣ activity in plants. Phosphorylation at some sites may serve to promote translation initiation, whereas phosphorylation of other sites may serve to inhibit protein synthesis or alter those mRNAs selected for translation. In this respect, it should be noted that phosphorylation at Ser-51 in eIF2␣ inhibits factor activity in mammalian cells (reviewed in Ref. 39), whereas it affects message selection in yeast (e.g. GCN4) through alterations in its re-initiation activity (46,47). eIF2␤ is required for ribosomal start site selection (48), and phosphorylation of mammalian eIF2␤ stimulates eIF2 activity (49). eIF2␤ is multiply phosphorylated in wheat leaves (11) as it is during early seed development, where approximately 4 -5 isoforms were observed. During subsequent seed development however, eIF2␤ underwent progressive dephosphorylation until a single basic species remained at 45 DAF. The presence of phosphorylated eIF2␤ correlated with the high rate of protein synthetic activity characterizing the mid-development of the seed whereas dephosphorylation of eIF2␤ during late development correlated with a corresponding low level of protein synthesis, observations suggesting that phosphorylation of eIF2␤ may play a role in regulating translational activity.
In contrast to the dynamic changes in phosphorylation of eIF4B, eIF2␣, and eIF2␤, eIF4A remained little changed during seed development or germination. The only exception to this was during the earliest stage of seed development when the seed is dominated by the nucellus, a maternal tissue that undergoes programmed cell death. During this developmental stage, 50% of the total eIF4A in the seed was phosphorylated, similar to the observations made in wheat leaves following a heat shock (11) or in maize roots following hypoxia (34). Therefore, phosphorylation of eIF4A appears to occur in response to stress and as part of some types of programmed cell death. During this same early developmental period, eIF4B was subject to degradation and eIF2␣ was present as phosphorylated intermediates, which may reflect the fate of these factors in the nucellus and other tissues as they undergo rapid programmed cell death.
Our observations suggest that the phosphorylation state of eIF4B, eIF2␣, eIF2␤, and even eIF4A is developmentally regulated but that the regulation of each exhibits some unique aspects. These differences may play a role in the changing translational environment of the developing seed and during acquisition of photosynthetic competence following germination as the complement of mRNAs can change dramatically from one developmental stage to another. Whether the translational requirements of mRNAs expressed only during specific developmental stages differs is unknown. However, changes in the phosphorylation status of the protein synthetic machinery may be a means by which necessary changes in the translational environment are achieved during plant development.