Glucose stimulated protein synthesis in pancreatic β -cells parallels an increase in the availability of the translational ternary complex (eIF2:GTP.Met:tRNAi) and the dephosphorylation of eIF2 α .

glucose in the presence of [ 35 S]-methionine for the final hour of treatment. Islets were also treated with 1 µ M thapsigargin (at 20mM glucose) as a control. Luciferase was immunoprecipitated from the cell lysates and separated on SDS-PAGE and the incorporation of [ 35 S]-methionine into luciferase was visualised by autoradiography (IP:Luciferase). Identical total cell lysates were also separated on a SDS-PAGE and Western blotted (WB) using anti-phospho-eIF2 α (Ser51) and anti-Erk2 antisera as a loading control. Results presented are representative of 3 separate experiments.


Summary
In pancreatic β-cells, glucose causes a rapid increase in the rate of protein synthesis. However, the mechanism by which this occurs is poorly understood. In this report, we demonstrate, in the pancreatic β-cell line MIN6, that glucose stimulates the recruitment of ribosomes onto the mRNA, indicative of an increase in the rate of the initiation step of protein synthesis. This increase in the rate of initiation is not mediated through an increase in the availability of the initiation complex eIF4F, as glucose is unable to stimulate eIF4F assembly or, in the absence of amino acids, modulate the phosphorylation status of 4E-BP1. Moreover, in MIN6 cells and isolated islets of Langerhans, rapamycin, an inhibitor of mTOR, only partially inhibited glucose stimulated protein synthesis.
However, we show that glucose stimulates the dephosphorylation of eIF2α in MIN6 cells and the assembly of the translational ternary complex, eIF2:GTP.Met:tRNAi, in both MIN6 cells and islets of Langerhans. The changes in the phosphorylation of eIF2α are not mediated by the PKR-like endoplasmic reticulum eIF2α kinase, PERK, as PERK is not phosphorylated at low glucose concentrations and over-expression of a dominant negative form of PERK has no significant effect on either glucose stimulated protein synthesis or the phosphorylation of eIF2α.
Taken together, these results indicate that glucose stimulated protein synthesis in pancreatic β-cells is regulated by a mechanism largely independent of the activity of mTOR but which is likely to be dependent on the availability of the translational ternary complex, regulated by the phosphorylation status of eIF2α.

Introduction.
The pancreatic ß-cell releases insulin in response to glucose. Additionally, there is a rapid two-fold increase in total protein synthesis, a 20 fold increase in proinsulin synthesis and the co-ordinate increase in the synthesis of a small subset of proteins, including the proinsulin endopeptidases PC2 and PC3 (1)(2)(3)(4)(5)(6)(7)(8)(9). These rapid increases in protein synthesis are almost entirely mediated at the translational level (1)(2)(3)(4)6,7,10) and ensure the replenishment of secreted proteins and the maintenance of ß-cell mass. However, the mechanism by which glucose regulates protein synthesis in pancreatic β-cells is poorly understood.
One key step in the regulation of protein synthesis is the recruitment of the 40S ribosome to the mRNA. This is promoted by the initiation complex, eIF4F, composed of the initiation factors eIF4G, a large scaffolding protein, eIF4E, the protein that binds to the 5'-cap structure (7mGTP) of the mRNA, and eIF4A, an RNA helicase (11). An important mechanism regulating eIF4F assembly involves a family of eIF4E binding proteins (4E-BPs, archetype 4E-BP1) (12,13). In their hypophosphorylated forms, 4E-BP1 competes with eIF4G for binding to eIF4E and can therefore inhibit eIF4F complex formation. Phosphorylation of 4E-BP1 leads to its dissociation from eIF4E, therefore allowing eIF4E to bind eIF4G, potentially leading to an increase in mRNA translation (14-determined by the Bradford assay (Bio-Rad) using BSA as standard. The protein lysates were stored at -80°C until further analysis.
Sedimentation Gradients and Polysome Analysis. MIN6 cells were incubated for 1h in either KRB or DMEM supplemented with 2mM glucose prior to the cells being incubated for a further 1h in either KRB or DMEM supplemented with either 2 or 20mM glucose. Cycloheximide was added for the last 5min to a final concentration of 0.1mg/ml. After treatment, the cells were scraped into polysome extraction buffer (1% Triton X-100, 20mM HEPES pH7.6, 2mM MgCl 2 , 130mM KCl, 1mg/ml heparin, 1mg/ml cycloheximide, 1mM DTT, 1µl/ml RNAguard) and lysed using 15 strokes of a potter homogenisor. Cell lysates were centrifuged for 10min at 13,000g at 4 0 C to remove nuclei and cellular debris. The supernatants were then layered onto 7 to 47% sucrose gradients (sucrose, 20mM HEPES pH7.6, 2mM MgCl 2 , 130mM KCl, 1mg/ml heparin, 1mg/ml cyclohexamide, 1mM DTT) and centrifuged for 2h at 39,000 rpm at 4°C in a TH64.1 rotor (Sorvall). The gradients were pumped, by upward displacement, through a UV monitor and the absorbance read at 254nm using an ISCO gradient collector.
Islet Isolation and Culturing. Pancreatic islets were isolated from 200-250g female Wistar rats by collagenase digestion and Histopaque density gradient centrifugation as previously described (3).
Islets were either used immediately or were cultured in RPMI 1640 containing 5.6mM glucose with Adenovirus Generation and Adenoviral Infection. cDNA encoding the firefly luciferase reporter downstream of the 5'UTR of ATF4 and cDNA encoding amino acids 1-583 of PERK (PERK∆C) were subcloned from the plasmids TK::ATF4UTR::LUC (30) and PERK.WT.9E10.pCDNA.amp (29) respectively, kindly provided by Prof. David Ron (Skirball Institute, New York University School of Medicine), into the shuttle vector pAd-Track-CMV prior to the generation of recombinant adenoviruses. Recombinant adenoviruses expressing ATF4 luciferase (Ad-ATF4luc) and PERK∆C (Ad-PERK∆C) were then prepared using the Ad-easy system (33) as previously described (34 To infect MIN6 cells or islets of Langerhans, the recombinant adenoviruses were mixed with culture medium and cells were exposed to the viruses with 100-200 multiplicity of infection for 24-48h prior to the experiment. Under these conditions, >90% of cells were infected, as determined by EGFP expression. Protein Synthesis Measurements. Protein synthesis measurements were performed as described previously (35). In these experiments, the cell treatments were performed in the presence of [ 35 S]methionine for 1h. All treatments were stopped by washing the cells with ice-cold PBS and lysing them in ice-cold extraction buffer (as described in Cell Culture and Treatment). Cell lysates were prepared and equal amounts of protein (about 20µg) were spotted onto 3MM Whatmann filters. The filters were then washed by boiling them for 1 min in 5% trichloroacetic acid containing 0.1g/litre of methionine. This was repeated three times. The filters were then dried and immersed in scintillant before determining radioactivity by scintillation counting.
The bound antibody/protein G complexes were then added to [ 35 S]-Methionine labelled MIN6 protein lysates and rotated for 2h at 4 0 C. The beads were then washed three times in RIPA buffer and the bound proteins finally eluted into Laemmli sample buffer (36) prior to being separated by

Glucose Stimulated Protein Synthesis is Mediated Through an Increase in the Rate of Initiation.
To investigate the mechanism by which glucose regulates protein synthesis in pancreatic β-cells, we initially determined the rates of protein synthesis at either low or high glucose concentrations in the pancreatic ß-cell line MIN6, a cell line that synthesises and secretes insulin in response to physiologically relevant glucose concentrations (6,38). In order to establish which step in protein synthesis may be important in glucose stimulated protein synthesis, we determined the polysome profile of MIN6 cells incubated at both 2 and 20mM glucose in either KRB or DMEM (Fig. 2). In cells incubated in either KRB or DMEM, increasing glucose concentration from 2 to 20mM led to a decrease in the 80S monosomal peak along with an increase in the number of polysomes, indicative of an increase in the rate of the initiation phase of protein synthesis. Additionally, the polysome to monosome ratio at either low or high glucose concentrations is greater when cells are incubated in DMEM compared to those incubated in KRB.
This indicates that the basal rates of protein synthesis are higher in cells incubated in DMEM than those incubated in KRB.

The Role of mTOR in Glucose Stimulated Protein Synthesis in MIN6 Cells. Amino acids and
glucose have been shown to influence the rate of protein synthesis in a number of cell types through the modulation of the phosphorylation status of 4E-BP1 and p70 S6 kinase via a mechanism that is dependent on mTOR. We therefore investigated the role of mTOR in glucose stimulated protein synthesis in MIN6 cells. MIN6 cells were pre-incubated for 1h in either KRB or DMEM supplemented with 2mM glucose and then further incubated for 1h in the same medium (KRB or DMEM) at either 2mM or 20mM glucose in the presence or absence of rapamycin, a specific inhibitor of mTOR (Fig. 3). Increasing glucose concentration in cells incubated in KRB (i.e. in the absence of amino acids) had no effect on the phosphorylation of either p70 S6 kinase or 4E-BP1 as assessed by determining: 1.) the mobility of both 4E-BP1 and p70 S6 kinase on SDS-PAGE; 2.) the phosphorylation of specific residues within 4E-BP1, Thr36/45 and Ser65, and 3.) the phosphorylation of ribosomal protein S6, a substrate for p70 S6 kinase (Fig. 3a). In contrast, in MIN6 cells incubated in DMEM, an increase in glucose concentration led to: 1) an increase in the phosphorylation of both 4E-BP1 and p70 S6 kinase, demonstrated by their decrease in mobility on SDS-PAGE; 2.) the phosphorylation of residues Thr36/45 and Ser65 within 4E-BP1, and 3.) an increase in the phosphorylation of ribosomal protein S6 at Ser235/236 (Fig. 3a). In this latter condition (i.e. in the presence of amino acids), rapamycin completely blocked both the basal and glucose stimulated phosphorylation of p70 S6 kinase, 4E-BP1 and S6, indicating that mTOR plays an important role in these events (Fig. 3a). Indeed, it has been previously demonstrated that amino acids are required for glucose stimulated phosphorylation of downstream targets of mTOR in the pancreatic ß-cell line RINm5F cells and in islets (39), and this explains why glucose cannot stimulate the phosphorylation of the downstream targets of mTOR in KRB.
As the phosphorylation of 4E-BP1 can lead to its release from eIF4E, an increase in eIF4F complex assembly (i.e. an increase in the association of eIF4G and eIF4A with eIF4E) and hence the rate of initiation, we therefore examined the effect of glucose on eIF4F assembly by pulling down eIF4E and associated proteins using 7mGTP Sepharose beads (Fig. 3b). In KRB (i.e. in the absence of amino acids), an increase in glucose concentration from 2 to 20mM had no detectable effect on the dissociation of 4E-BP1 from eIF4E, on eIF4F assembly, or on the association of PABP with the eIF4F complex. Additionally, treatment of cells with rapamycin had no significant effect on glucose stimulated protein synthesis in KRB (Fig. 3c). Therefore, the activation of downstream targets of mTOR and the stimulation of eIF4F assembly play no detectable role in glucose stimulated protein synthesis in KRB (i.e. in the absence of amino acids). In DMEM, basal levels of the eIF4F complex and the levels of associated PABP were greater than in KRB. An increase in glucose concentration from 2 to 20mM in DMEM caused a moderate release of 4E-BP1 from eIF4E. However, glucose had no detectable effect on eIF4F assembly or the association of PABP with the eIF4F complex ( Fig. 3b). Treatment of cells with rapamycin led to an increase in the association of 4E-BP1 with eIF4E but no dissociation of the eIF4F complex. Additionally, rapamycin inhibited glucose stimulated protein synthesis by only 26% in DMEM (i.e. in the presence of amino acids) (Fig. 3c).
Therefore, the activation of downstream targets of mTOR and the stimulation of eIF4F assembly is not the major determinant in glucose stimulated protein synthesis in DMEM (i.e. in the presence of amino acids).

Glucose Stimulates the Rapid, Dose Dependent Dephosphorylation of eIF2α. One important and
well characterised mechanism by which the initiation of translation and hence protein synthesis is regulated is via the phosphorylation of the alpha subunit of eIF2 (eIF2α) at Ser51.
We therefore investigated the phosphorylation of eIF2α in response to glucose in MIN6 cells. MIN6 cells were pre-incubated for 1h in either KRB or DMEM supplemented with 2mM glucose and then further incubated for 1h in the same medium (i.e. KRB or DMEM) at either 2mM or 20mM glucose ( Fig. 4a). Using a phospho-specific antibody to eIF2α Ser51, we show that eIF2α is phosphorylated at low glucose concentration (Fig. 4a). However, at high glucose concentration, eIF2α is dephosphorylated (Fig. 4a). The dephosphorylation of eIF2α was shown to occur between 2 and 10mM, within the physiological range of glucose concentrations (Fig. 4bi). Additionally, decreases in the phosphorylation of eIF2α in response to glucose paralleled increases in the rates of protein synthesis (Fig. 4bii).
The temporal kinetics of the regulation of eIF2α phosphorylation in response to glucose were also investigated ( Fig. 4c). Transfer of cells from DMEM complete media containing 25mM glucose into KRB supplemented with 2mM glucose led to the phosphorylation of eIF2α within 20min. After a 1h incubation in KRB supplemented with 2mM glucose, the cells were then further incubated with KRB supplemented with 20mM glucose. This led to a rapid (within 15 minutes) dephosphorylation of eIF2α (Fig. 4c).

Glucose Stimulates the Assembly of the Translational Ternary Complex.
A likely consequence of eIF2α dephosphorylation is an increase in the assembly of the translational ternary complex via an increase in the activity of eIF2B. In order to determine whether ternary complex formation is stimulated by glucose in vivo, we monitored the expression of luciferase in cells infected with recombinant adenovirus encoding luciferase downstream of the 5'UTR of ATF4 and under the control of a CMV promoter (Ad-ATF4Luc). ATF4 is a transcription factor whose translation is regulated by short open reading frames (uORFs) in its 5'UTR by a mechanism analogous to that which regulates GCN4 expression in yeast (30). Briefly, upon an increase in ternary complex availability, ATF4 expression is repressed. However, when the availability of ternary complex is low, ATF4 expression is up-regulated (30).  (Fig. 5). We also investigated eIF2α phosphorylation in identical samples used to investigate changes in ternary complex formation using a phospho-specific antibody to eIF2α Ser51 (Fig. 5). In the presence of thapsigargin and at low glucose concentrations, either in the presence or absence of amino acids, eIF2α is phosphorylated. However, upon an increase in glucose concentration, eIF2α is dephosphorylated. Taken together, these results indicate that glucose stimulates ternary complex formation in MIN6 cells incubated in either KRB or DMEM and that these changes parallel the dephosphorylation of eIF2α.

Role of PERK in glucose stimulated protein synthesis.
In the pancreatic ß-cell line MIN6, eIF2α phosphorylation is regulated in response to changes in glucose concentration. At low glucose concentration (2mM), eIF2α is phosphorylated, whereas at high glucose concentration (20mM), eIF2α is dephosphorylated. These changes in the phosphorylation of eIF2α parallel changes in both the rate of protein synthesis and changes in ternary complex formation. However, the kinase responsible for the phosphorylation of eIF2α upon glucose deprivation is unknown. One possible candidate is the eIF2α kinase PERK (29,40).
Upon activation, PERK is auto-phosphorylated at Thr980. Therefore, we investigated the phosphorylation status of PERK in response to glucose in MIN6 cells using a phospho specific antibody to PERK (Thr980). MIN6 cells were pre-incubated for 1h in KRB supplemented with 0.5mM glucose and then further incubated for 1h in the same medium at either 0.5mM glucose, 20mM glucose, or 20mM glucose plus thapsigargin as control (Fig. 6a). As expected, treatment of MIN6 cells with thapsigargin led to a large increase in the phosphorylation of PERK (Fig. 6a).
However, no significant change in the phosphorylation of PERK was detected between low (0.5mM) and high (20mM) glucose concentrations (Fig. 6a).
As a corollary of the above, MIN6 cells were mock infected or infected with recombinant adenoviruses over-expressing either a dominant negative mutant of PERK (Ad-PERK∆C) (29) or EGFP (Ad-emptyEGFP) only as control. 24h post-infection, the cells were incubated in KRB containing 0.5mM glucose for 1h prior to incubation in the same buffer at either 0.5mM or 20mM glucose for a further 1h. MIN6 cells were also treated with thapsigargin as control. Following treatment, the phosphorylation status of eIF2α (Fig. 6bii) and the rates of protein synthesis (Fig.   6bi) were determined. Over-expression of the dominant negative PERK inhibited thapsigargin induced eIF2α phosphorylation, and blocked a significant proportion of the inhibitory effect of thapsigargin on the rate of protein synthesis (Fig. 6b). However, expression of the dominant negative PERK had only a partial effect on the phosphorylation of eIF2α at low glucose and had no effect on glucose stimulated protein synthesis (Fig 6b).

Glucose Stimulated Protein Synthesis is Independent of mTOR and Parallels an Increase in
Ternary Complex Formation in islets of Langerhans. In order to confirm whether the results obtained in MIN6 cells also occur in primary β-cells, islets of Langerhans were isolated from rats and pre-incubated for 2h in KRB or DMEM supplemented with 2mM glucose and then further incubated in the same medium (i.e. KRB or DMEM) for 1h at either 2mM or 20mM glucose in the presence or absence of rapamycin. Increasing glucose concentration from 2 to 20mM led to a significant increase in [ 35 S]-methionine incorporation into total protein in islets incubated in either KRB or DMEM (Fig. 7ai). Additionally, glucose stimulated increases in the rate of protein synthesis were only partially affected in KRB and unaffected in DMEM by the addition of rapamycin (Fig. 7ai). As a control, the phosphorylation of ribosomal protein S6 was also monitored using a phospho specific antibody to S6 (Ser 235/236) (Fig. 7aii). As in MIN6 cells, S6 phosphorylation was stimulated at 20mM glucose in DMEM but not in KRB and glucose stimulated phosphorylation of S6 was completely inhibited by rapamycin (Fig. 7aii).
To investigate changes in ternary complex availability, rat islets of Langerhans were infected 48h with the Ad-ATF4luc adenovirus. 48h post-infection, the islets were pre-incubated for 2h in KRB  (Fig. 7b). The phosphorylation status of eIF2α was also investigated using a phospho-specific antibody to eIF2α(ser51). However, no consistent change in eIF2α phosphorylation was observed in response to either glucose or thapsigargin (Fig.   7b).

Discussion
Pancreatic β-cells are almost unique in their ability to sense and respond to changes in extracellular glucose, with increases in glucose concentration leading not only to rapid increases in proinsulin synthesis, but also to a rapid increase in the rate of total protein synthesis. This short-term glucose stimulated protein synthesis (within 1h) is almost entirely regulated at the post-transcriptional level through an increase in the rate of translation (1,2,6,7) (and this report, Fig. 1). However, the mechanism(s) by which glucose stimulates an increase in protein synthesis in pancreatic β-cells is poorly understood.
In this report, we demonstrate, in the pancreatic β-cell line MIN6, that glucose stimulates protein synthesis in the presence or absence of amino acids and that these increases are likely primarily mediated through an increase in the rate of initiation of translation (figure 2). These increases in protein synthesis occur via a mechanism that is independent of the activation of mTOR in the absence of amino acids and only partially dependent on mTOR in the presence of amino acids. In contrast, we show that glucose stimulates both the assembly of the translational ternary complex, eIF2.GTP:Met-tRNA i and the dephosphorylation of eIF2α, both in the absence or presence of amino acids. Importantly, these changes parallel an increase in the rate of protein synthesis in response to glucose. The role of mTOR in glucose stimulated protein synthesis and glucose stimulated changes in ternary complex formation were confirmed in islets. Therefore, taken together, our data show that eIF2α phosphorylation is likely to be key in regulating glucose stimulated protein synthesis in pancreatic β-cells. However, we cannot definitively conclude that the results obtained in MIN6 cells or islets occur in primary β-cells.
It has been reported that glucose stimulates the de-phosphorylation of eEF2 in both islets and MIN6 cells between 1mM and 10mM, indicating that glucose may also stimulate an increase in the rate of elongation (41). Indeed, it has also been reported that the rate of protein elongation is specifically up-regulated between 0mM-3.3mM glucose in islets (42). However, these increases in the elongation rate are fairly modest and occur only at these low non-physiological glucose concentrations (42). Hence, these changes are not thought to play a major role in glucose stimulated protein synthesis. Therefore, it is likely that the primary mechanism by which glucose stimulates protein synthesis in pancreatic β-cells is through an increase in the rate of initiation.
In a number of other cell types, nutrients such as glucose or amino acids have been shown to stimulate protein synthesis. In many of these cases, this has been linked to an increase in the activity of the protein kinase mTOR (for reviews see (43-45)) which is required for the phosphorylation of both 4E-BP1 and p70 S6 kinase, two important modulators of protein synthesis. Hyperphosphorylation of 4E-BP1 stimulates its dissociation from eIF4E and the assembly of the eIF4F complex and hence the rate of initiation from cap-dependent mRNAs (14)(15)(16). The phosphorylation and activation of p70 S6 kinase leads to the phosphorylation of ribosomal protein S6, which has been reported to lead to an increase in the rate of initiation of a specific subset of mRNAs, which have a tract of polypyrimidines in their 5'UTR (46-48). Indeed, in this report, we demonstrate in MIN6 cells that glucose, in the presence of amino acids, stimulates the phosphorylation of both 4E-BP1 and p70 S6 kinase through a mechanism dependent on mTOR (Fig.3). This is in agreement with previous reports demonstrating that amino acids are required for glucose stimulated phosphorylation of 4E-BP1 and p70 S6 kinase in islets and in the pancreatic β-cell line RINm5F (9,39). We additionally demonstrate that glucose stimulates the phosphorylation of 4E-BP1 on amino acid residues Ser65 and Thr36/45 (Fig. 3a). Importantly, the phosphorylation of these residues is known to be critical in stimulating the release of 4E-BP1 from eIF4E (49,50). Accordingly, we also show that glucose stimulates the release of 4E-BP1 from eIF4E (Fig. 3).
Surprisingly, these changes do not lead to any detectable increase in eIF4F assembly (Fig. 3b).
Therefore, the functional significance of the phosphorylation of the downstream targets of mTOR is unclear. However, treatment of MIN6 cells or islets with rapamycin, a specific inhibitor of mTOR, did lead to a partial inhibition of glucose stimulated protein synthesis (Figs. 3c and 7a). In the absence of amino acids, under conditions where there is no detectable phosphorylation of the downstream targets of mTOR or any changes in eIF4F assembly, glucose can still stimulate protein synthesis in both MIN6 cells and islets (Figs 3 and 7a). Furthermore, this glucose stimulated protein synthesis is unaffected by the addition of rapamycin in MIN6 cells and only partially affected in islets (Figs 3 and 7a). Clearly, these data show that glucose can stimulate protein synthesis through a mechanism independent of mTOR. Therefore, we believe that glucose stimulated protein synthesis, either in the presence or absence of amino acids, is likely to be primarily regulated by a mechanism independent of the activation of mTOR.
An alternative mechanism by which the rate of initiation and hence protein synthesis could potentially be up-regulated in response to glucose is through an increase in the availability of the translational ternary complex. Interestingly, we demonstrate, both in islets and MIN6 cells, that the availability of the translational ternary complex is up-regulated at high glucose concentrations in vivo, either in the presence or absence of amino acids (Figs. 5 and 7b). It is likely that glucose stimulates the assembly of the translational complex, at least in part, through the activation of the guanine nucleotide exchange factor eIF2B, as eIF2B activity is transiently increased in response to glucose in rat islets (20). The mechanism by which eIF2B is regulated by glucose is unknown but one well characterised mechanism by which eIF2B can be regulated is through the phosphorylation of eIF2α on ser51. In this report, we demonstrate that eIF2α becomes rapidly dephosphorylated in response to increasing glucose concentrations in MIN6 cells and that these changes in eIF2α phosphorylation parallel changes in glucose stimulated protein synthesis ( Fig. 4a and b). Moreover, glucose stimulated dephosphorylation of eIF2α occurs in both DMEM and KRB (Fig. 4a). In contrast, in islets, we couldn't detect any consistent change in the phosphorylation status of eIF2α in response to glucose (in agreement with Gilligan et al. (20)) (Fig. 7b). Additionally, we couldn't detect any increase in the phosphorylation of eIF2α after treatment with thapsigargin (an agent known to induce eIF2α phosphorylation) (Fig 7b). However, increased glucose concentration clearly led to an increase in the assembly of the translational ternary complex whereas thapsigargin treatment resulted in a large inhibition of ternary complex formation (Fig. 7b). One potential reason that could explain why we cannot detect changes in eIF2α phosphorylation in islets (even after treatment with thapsigargin) is that phosphorylated eIF2α in non-β-cells, or perhaps much more likely, highly phosphorylated eIF2α in a population of stressed cells, may mask the effects of glucose or thapsigargin on the phosphorylation of eIF2α in β-cells. Indeed, cells at the centre of cultured islets can suffer from hypoxia, a stress known to induce eIF2α phosphorylation in other cell types (51). Therefore, given that glucose stimulates the dephosphorylation of eIF2α in MIN6 cells, we believe that it is likely that glucose stimulated assembly of the translational ternary complex in islets is also mediated by the dephosphorylation of eIF2α.
Two regulatory subunits of protein phosphatase 1 have been implicated in eIF2α dephosphorylation.
These include GADD34, whose expression is upregulated in response to cell stress (i.e. after a few hours) in order to stimulate the recovery from the shut-off of protein synthesis (52,53) and CReP, which is constitutively active and acts to maintain the basal phosphorylation state of eIF2α (54).
Given the temporal kinetics of eIF2α dephosphorylation in response to glucose in MIN6 cells (Fig.   4c), it is unlikely that this event is regulated through the up-regulation of GADD34 expression. We believe it is more likely that these changes in phosphorylation of eIF2α are regulated through the activity of an eIF2 kinase in combination with a constitutively active phosphatase.
Glucose starvation is known to stimulate eIF2α phosphorylation in C6 glial cells (55) and whole mouse pancreata (56). It is also known that glucose starvation can cause endoplasmic reticulum stress, hence it was previously speculated that PERK may phosphorylate eIF2α in response to glucose deprivation (57). However, we show that over-expression of a dominant negative mutant of PERK (Ad-PERK∆C) (29) in MIN6 cells had only a partial effect on the phosphorylation of eIF2α at either low or high glucose concentrations and had no effect on glucose stimulated protein synthesis (Fig. 6b). Additionally, no increase in the phosphorylation of PERK was detected at low glucose compared to high glucose (Fig. 6a). These data demonstrate that PERK is unlikely to be the primary eIF2 kinase responsible for eIF2α phosphorylation in response to low glucose concentration in pancreatic β-cells. However, the eIF2 kinase mGCN2 is regulated by nutrients. For example, mGCN2 is activated by amino acid starvation (26). Additionally, its yeast orthologue, GCN2, is also activated in response to glucose starvation (27). However, it has not been demonstrated that mGCN2 activity is regulated by glucose in mammalian cells. Nevertheless, it is possible that mGCN2 is the kinase responsible for the phosphorylation of eIF2α in pancreatic β-cells under conditions of low glucose concentration and this will be the subject of further investigation.