Ser-64 and Ser-111 in PHAS-I Are Dispensable for Insulin-stimulated Dissociation from eIF4E*

Insulin stimulates phosphorylation of multiple sites in the eIF4E-binding protein, PHAS-I, leading to dissociation of the PHAS-I·eIF4E complex and to an increase in cap-dependent translation. The Ser-64 and Ser-111 sites have been proposed to have key roles in controlling the association of PHAS-I and eIF4E. To determine whether the effects of insulin require these sites, we assessed the control of PHAS-I proteins having Ala-64 or Ala-111 mutations. The results indicate that phosphorylation of neither site is required for insulin to promote release of PHAS-I from eIF4E. Also, the mutation of Ser-111, which has been proposed to serve as a necessary priming site for the phosphorylation of other sites in PHAS-I, did not affect the phosphorylation of Thr-36/45, Ser-64, or Thr-69. Insulin promoted the release of eIF4E from PHAS-II, a PHAS isoform that lacks the Ser-111 site, but it was without effect on the amount of eIF4E bound to the third isoform, PHAS-III. The results demonstrate that contrary to widely accepted models, Ser-64 and Ser-111 are not required for the control of PHAS-I binding to eIF4E in cells, implicating phosphorylation of the Thr sites in dissociation of the PHAS-I·eIF4E complex. The findings also indicate that PHAS-II, but not PHAS-III, contributes to the control of protein synthesis by insulin.

Insulin and certain amino acids stimulate cap-dependent translation by promoting the phosphorylation of PHAS-I, the best characterized member of a family of three translational repressor proteins expressed in a wide variety of cell types (1,2). Hypophosphorylated PHAS-I binds tightly to eIF4E 1 (3,4), the mRNA cap-binding protein, and inhibits cap-dependent translation by blocking the association between eIF4E and eIF4G (5,6). When phosphorylated in the appropriate sites, PHAS-I dissociates from eIF4E (3,4), allowing the formation of the 5Ј complex needed for efficient binding and/or scanning by the 40 S ribosomal subunit. The signaling pathways utilized by insulin and amino acids converge at the level of mTOR (1,2). Thus, rapamycin and other inhibitors of mTOR attenuate the phosphorylation of PHAS-I.
The following six phosphorylation sites in PHAS-I have been identified: Thr-36; Thr-45; Ser-64; Thr-69; Ser-82; and Ser-111 (7,8). Which sites are most important in the control of the association of PHAS-I and eIF4E in cells has not been established. The sites conform to a (Ser/Thr)-Pro motif (7) with the exception of Ser-111, which is followed by Gln. Thr-36, Thr-45, Ser-64, and Thr-69 are phosphorylated in response to insulin (7,9,10), and these sites are conserved in PHAS-I proteins from different species as well as in the two other PHAS isoforms (2). The Ser-111 site is found in mammalian PHAS-I proteins, but not in the PHAS-I proteins from other vertebrates such as chicken or fish, or in mammalian PHAS-II or PHAS-III (2). Ser-82 is less conserved than the other (Ser/Thr)-Pro sites, and phosphorylation of Ser-82 in vitro does not inhibit binding to eIF4E (11). Moreover, in cells, this site appears to contain relatively little phosphate and it is insensitive to insulin (7,9). Therefore, Ser-82 does not appear to contribute significantly to the control of PHAS-I. There is a great deal of interest in the phosphorylation of the other two Ser phosphorylation sites, Ser-64 and Ser-111. Ser-64 undergoes the most dramatic increase in phosphorylation in response to insulin (12), and it is phosphorylated only after phosphorylation of the three Thr-Pro sites (10,13). Ser-64 has been considered to be a prime candidate for promoting dissociation of the PHAS-I⅐eIF4E complex, because its phosphorylation in vitro markedly decreases the affinity of PHAS-I for eIF4E (3,9,14). Ser-111 was first identified as the major site for in vitro phosphorylation by protein kinase CK2 (15). More recently, this site has been shown to be phosphorylated in vitro by the phosphatidylinositol 3-OH kinase-related protein kinases, ATM, ATR, and Smg1 (15)(16)(17)(18). Phosphorylation of Ser-111 has been suggested to serve as a priming event, which is necessary for both the phosphorylation of other sites in PHAS-I and the dissociation of the PHAS-I⅐eIF4E complex (8,18). This study was conducted to investigate the role of Ser-64 and Ser-111 in the control of PHAS-I binding to eIF4E.

EXPERIMENTAL PROCEDURES
Antibodies-Antibodies to the COOH-terminal regions of PHAS-I and eIF4E (19) and the phosphospecific antibodies, P-Thr-36/45, P-Ser-64, and P-Thr-69 (10), were generated in rabbits and affinity-purified as described previously (10,19). The sequences of amino acids immediately surrounding Thr-36 and Thr-45 are identical. Consequently, the P-Thr-36/45 antibodies bind to PHAS-I phosphorylated in either Thr-36 or Thr-45 (10). Monoclonal antibody to the FLAG epitope tag was from Sigma. Monoclonal antibodies 9E10 and 12CA5, which recognize the Myc and HA epitope tags, respectfully, were purified from hybridoma culture medium.
A construct (pKH3 PHAS-III ) for expressing HA-tagged PHAS-III in cells was generated by excising PHAS-III cDNA from pGEX-2T with BamHI and EcoRI and inserting it between these sites in pKH3 (20). Human PHAS-III in pGEX-2T was supplied by Dr. Tai-An Lin.
To prepare an expression vector for FLAG-tagged eIF4E (pCMV-Tag 2B eIF4E ), EcoRI sites were added to the ends of human eIF4E cDNA by using the polymerase chain reaction with 5Ј-GGAATTCATGGCGACT-GTCGAACCGG-3Ј and 5Ј-GGAATTCTTAAACAACAAACCTATTTTT-AG-3Ј as forward and reverse primers, respectively. The product was digested with EcoRI and inserted into the EcoRI site of pCMV-Tag 2B (Stratagene). PHAS and eIF4E clones having the proper orientation were selected by restriction mapping. Inserts encoding the proteins were sequenced and found to be free of errors. 2 Cell Culture and Transfections-HEK293 cells (2.5 ϫ 10 6 ) were seeded into 150-mm dishes and cultured for 24 h at 37°C in Dulbecco's modified Eagle's medium supplemented with 2 mM L-glutamine, 10% fetal bovine serum, 10 units/ml penicillin G, and 10 g/ml streptomycin. A modified CaPO 4 procedure was used to introduce DNA (34 g each of the PHAS-I and eIF4E expression constructs or the pCMV-Tag 2B or pCMV-Tag 3A vectors) into the cells (21). Cells were incubated with the precipitated DNA for 4 h. The medium was then replaced, and the cells were cultured in growth medium. After 15 h, cells from each 150-mm dish were removed using 0.02% trypsin, seeded into six 60-mm-diameter culture dishes, and cultured for 36 h. This procedure provided equal expression of proteins among treatment groups. The cells were serumstarved prior to incubations with insulin and/or rapamycin. For serum starvation, the growth medium was replaced with Dulbecco's modified Eagle's medium containing 0.2% bovine serum albumin (Intergen) and the cells were incubated at 37°C for 15 h. The Dulbecco's modified Eagle's medium was replaced with buffer (145 mM NaCl, 5.4 mM KCl, 1.4 mM CaCl 2 , 1.4 mM MgSO 4 , 5 mM glucose, 0.5% bovine serum albumin, 0.1 mM sodium phosphate, 25 mM NaHCO 3 , and 10 mM HEPES, pH 7.4), and the cells were incubated for 1.5 h. Cells were incubated with or without rapamycin (20 nM) for 1 h and then incubated for an additional 30 min after adding insulin (100 milliunits/ml) and/or amino acids (2.5ϫ minimum essential medium amino acid mixture). To ter-minate the incubations, the cells were placed on ice, rinsed with chilled lysis buffer minus Nonidet P-40, and then immediately scraped into 1 ml of lysis buffer (100 mM NaCl, 50 mM NaF, 10 mM EDTA, 1% Nonidet P-40, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 500 M microcystin, 10 mM sodium pyrophosphate, and 50 mM HEPES, pH 7.4). The samples were mixed at 4°C for 1 h and centrifuged at 13,000 ϫ g for 20 min before the supernatants were retained for analyses. Extract protein concentrations were adjusted to 0.5 mg/ml by adding lysis buffer. Protein was measured by using bicinchoninic acid (22).
Immunoprecipitation of PHAS-I⅐eIF4E Complexes-eIF4E proteins bound to epitope-tagged PHAS proteins were isolated by incubating extracts (0.5 ml) for 18 h at 4°C with protein G-agarose (Invitrogen) (10 l of packed beads), 0.2% IgG-free bovine serum albumin, and 5 g of the 9E10 monoclonal antibody for Myc-tagged PHAS-I and PHAS-II proteins or with 5 g of 12CA5 for HA-tagged PHAS-III. Immune complexes bound to the beads were recovered by centrifugation. The beads were washed twice (1 ml/wash) with lysis buffer and twice with lysis buffer minus Nonidet P-40. Proteins were eluted from the beads by using SDS sample buffer (23).
Affinity Purification of eIF4E Complexes-eIF4E was partially purified by using m 7 GTP-Sepharose 4B (Amersham Biosciences). Extract samples (250 -300 g protein) were mixed with 20 l of m 7 GTP-Sepharose 4B before the beads were washed essentially as described previously (3). Proteins were eluted from the beads with SDS sample buffer (23).
Electrophoretic Analyses-Samples were subjected to electrophoresis using the method of Laemmli (23). Proteins were transferred to membranes (Immobilon) and immunoblotted with the appropriate antibodies as described previously (19). Bound antibodies were detected by using alkaline phosphatase-conjugated secondary antibodies and CDP-Star (Tropix). Signal intensities were determined by scanning laser densitometry of films (Kodak X-Omat AR-5).

Influence of Ser-64 on the Control of PHAS-I by Insulin-To
determine whether Ser-64 phosphorylation was required for the control of PHAS-I by insulin, a Ser to Ala mutation was introduced in the site to prevent its phosphorylation in cells.  Myc epitope-tagged forms of this mutant protein and wild type PHAS-I were transiently overexpressed in HEK293 cells. Overexpressing the translational repressor, PHAS-I, may dramatically decrease cap-dependent translation by decreasing eIF4E availability (4). Because inhibition of protein synthesis may alter PHAS-I phosphorylation (24), which would complicate the analyses, we attempted to maintain eIF4E levels by coexpressing FLAG epitope-tagged eIF4E. The transfected cells were treated with insulin, amino acids, and rapamycin before the PHAS-I proteins were immunoprecipitated with Myc antibodies. Samples were then subjected to SDS-PAGE, and immunoblots were prepared with antibodies to either PHAS-I (Fig. 1A) or eIF4E ( Fig. 2A).
As described previously for endogenous PHAS-I (3), Myctagged wild type PHAS-I appeared in immunoblots as three bands designated ␣, ␤, and ␥ (Fig. 1A). No PHAS-I was detected in immunoprecipitates from the vector control cells, indicating that the signals detected were because of the overexpressed protein. Insulin decreased the ␣ form and increased the ␥ form (Fig. 1B). Rapamycin opposed these actions of insulin and promoted an increase in the ␣ form. Supplementing the medium with amino acids had little effect on the electrophoretic mobility of PHAS-I in either the absence or presence of insulin.
The mobility shifts promoted by insulin and rapamycin result from increases and decreases, respectively, in the phosphorylation of PHAS-I. The shift from ␤ to ␥ occurs with the phosphorylation of Ser-64 (2), which in the hierarchy of PHAS-I phosphorylation is the last site phosphorylated (10, 13). As expected, mutating Ser-64 to Ala blocked formation of ␥, even in insulin-treated cells (Fig. 1A). Although no ␥ form was present, insulin and rapamycin clearly affected the distribution of PHAS-I between the ␣ and ␤ forms. Indeed, when the ␣ form was expressed as a percentage of the respective PHAS-I protein immunoprecipitated, the mutant protein was indistinguishable from wild type (Fig. 1B).
Insulin decreased by ϳ50% the amounts of both FLAG-eIF4E and endogenous eIF4E that coimmunoprecipitated with wild type PHAS-I (Fig. 2, A and B). This result is consistent with the well established action of the hormone to promote dissociation of the PHAS -I⅐eIF4E complex (1, 2). Mutating Ser-64 to Ala did not attenuate the response to insulin, indicating that phosphorylation of Ser-64 is not required for dissociation of the complex. Indeed, the amount of eIF4E complexed with Ala-64 PHAS-I was indistinguishable from that bound to wild type PHAS-I after all of the treatments tested (Fig. 2B). Rapamycin attenuated the effects of insulin on increasing the dissociation of eIF4E from both the wild type and mutant PHAS-I proteins. Without insulin, amino acids slightly decreased the amount of eIF4E that coimmunoprecipitated with wild type PHAS-I protein but did not affect the amount of eIF4E that coimmunoprecipitated with Ala-64 PHAS-I. Amino acids did not significantly enhance the effect of insulin on promoting dissociation of eIF4E from either PHAS-I protein (Fig. 2B). Thus, as we have demonstrated previously (9,10), the effect of insulin on PHAS-I does not require the addition of amino acids to the medium.
Influence of Ser-111 on Control of PHAS-I-An approach similar to that described above for investigating the role of Ser-64 was used to determine whether the phosphorylation of Ser-111 was needed for the control of PHAS-I by insulin. Mutating Ser-111 to Ala was without effect on the electrophoretic mobility of PHAS-I in either the absence or presence of insulin, rapamycin, or amino acids (Fig. 3, A and B). Thus, the mobility shift assay provided no indication that Ser-111 phosphorylation was required for the phosphorylation of other sites in PHAS-I. However, because phosphorylation of some sites (Thr-36 and Thr-45, for example) has relatively little effect on the electrophoretic mobility of PHAS-I in SDS-PAGE (2), we investigated phosphorylation of different sites by immunoblotting with phosphospecific antibodies.
In the experiments presented in Fig. 4, wild type PHAS-I and Ala-111 PHAS-I were immunoprecipitated and probed with phosphospecific antibodies to the Thr-36/45, Thr-69, and Ser-64 sites (10). Insulin increased the reactivity of PHAS-I with all three phosphospecific antibodies (Fig. 4, A-D). As noted previously (9,10), the effect of insulin on Thr-36/45 phosphorylation (Fig. 4B) was less than the effects of the hormone on the phosphorylation of either Ser-64 (Fig. 4C) or Thr-69 (Fig. 4D). The phosphorylation of Ser-64 and Thr-69 was more sensitive to inhibition by rapamycin than the phosphorylation of Thr-36/45.
When corrected for protein expression, the reactivities of the three phosphospecific antibodies for wild type and Ala-111 PHAS-I were indistinguishable (Fig. 4, B-D), confirming that Ser-111 phosphorylation is not required for the phosphorylation of Thr-36/45, Ser-64, and Thr-69. Consistent with the lack of effect of the Ala-111 mutation on phosphorylation, eIF4E binding to Ala-111 PHAS-I assessed by coimmunoprecipitation was almost identical to binding of the initiation factor to wild type PHAS-I (Fig. 5, A and B).
Effect of Insulin and Rapamycin on PHAS-II and PHAS-III Binding to eIF4E-The five (Ser/Thr)-Pro sites found in PHAS-I are also present in PHAS-II; however, PHAS-II lacks the equivalent of the Ser-111 site (25). Therefore, we assessed the effects of insulin and rapamycin on the association of PHAS-II and eIF4E. Neither insulin nor rapamycin changed the electrophoretic mobility of PHAS-II (Fig. 6A). This was not unexpected because PHAS-II does not undergo the same shifts in electrophoretic mobility that PHAS-I does when it is phosphorylated (19). Treating cells with insulin decreased by ϳ40% the amount of both FLAG-eIF4E and endogenous eIF4E that coimmunoprecipitated with PHAS-II (Fig. 6B), indicative of dissociation of the PHAS-II⅐eIF4E complex. To confirm that insulin stimulated the dissociation of PHAS-II and eIF4E, we partially purified eIF4E by using m 7 GTP-Sepharose to estimate the amount of PHAS-II present in a cap-binding complex with eIF4E (Fig. 6C). Insulin decreased by ϳ40% the amount of PHAS-II that copurified with eIF4E. Amino acids did not affect the amount of eIF4E that coimmunoprecipitated with PHAS-II (Fig. 6B) or the amount of PHAS-II that copurified with eIF4E ( Fig. 6C) in either the absence or presence of insulin. Rapamycin attenuated but did not abolish the effects of insulin on promoting dissociation of the PHAS-II⅐eIF4E complex (Fig. 6, B  and C). Thus, the association of PHAS-II and eIF4E appears to be controlled by insulin in a manner similar to PHAS-I and eIF4E.
We next investigated the effects of insulin, amino acids, and rapamycin on PHAS-III. Similar to PHAS-II, this isoform binds eIF4E, lacks the Ser-111 site (2), and does not undergo a shift in electrophoretic mobility in response to insulin or rapamycin (Fig. 7A). In contrast to both PHAS-I and PHAS-II, PHAS-III did not undergo appreciable dissociation from eIF4E when cells were incubated with insulin (Fig. 7B). DISCUSSION Insulin-stimulated dissociation of the PHAS-I⅐eIF4E complex in cells has been extensively documented, but relatively little information exists on the control of the other two PHAS isoforms. Results of this study provide the first demonstration that insulin promotes the dissociation of the PHAS-II⅐eIF4E complex in cells (Fig. 6). The decrease in PHAS-II bound to eIF4E produced by insulin was comparable to the decrease in PHAS-I bound to eIF4E. In contrast, insulin did not decrease the amount of PHAS-III bound to the initiation factor (Fig. 7) (26). Thus, it seems that PHAS-I and PHAS-II, but not PHAS-III, are mediators of insulin action on protein synthesis.
Phosphospecific antibodies are not available for the sites in PHAS-II. For this reason, we did not investigate the effects of insulin on the phosphorylation of individual sites in this PHAS isoform. However, previous results in 32 P-labeled 3T3-L1 adipocytes indicate that insulin stimulates the phosphorylation of PHAS-II in a rapamycin-sensitive manner (19). This implies that the control of PHAS-II phosphorylation by insulin is similar to that of PHAS-I, although there must be some difference in the phosphorylation of the two proteins since PHAS-II contains Ala in the position equivalent to Ser-111 in PHAS-I. In view of the well established effect of phosphorylation on binding of PHAS-I to eIF4E, it seems reasonable to conclude that dissociation of the PHAS-II⅐eIF4E complex resulted from phosphorylation of PHAS-II.
Two motifs required for the efficient phosphorylation of PHAS-I in cells have been described previously (27,28). The RAIP motif (named for the sequence Arg-Ala-Ile-Pro) is found in the NH 2 -terminal regions of PHAS-I and PHAS-II, but it is absent in PHAS-III (28). The TOS (for mTOR signaling) motif is formed by the last five amino acids (Phe-Glu-Met-Asp-Ile) in all three PHAS isoforms (27). Mutations in either the TOS or RAIP motifs reduced the phosphorylation of overexpressed PHAS-I proteins (27,28). PHAS-I was recently shown to bind to the mTOR-associated protein, raptor (29, 30), which has been proposed to present substrates to the mTOR kinase (29). Mutating either domain also abolished tight binding of PHAS- I   FIG. 3. Effects of insulin, rapamycin, and amino acids on the electrophoretic mobilities of Myc-tagged wild type and Ala-111 PHAS-I. HEK293 cells were transfected with vectors alone (pCMV-Tag 2B and pCMV-Tag 3A) or with pCMV-Tag 2B eIF4E plus either pCMV-Tag 3A PHAS-I or pCMV-Tag 3A A111 PHAS-I . The cells were then incubated as described in the legend to Fig. 1. A, a PHAS-I blot is presented. B, the relative intensities of the ␣, ␤, and ␥ bands were determined by optical density scanning of films and expressed as percentages of the total. Mean values ϩ S.E. from seven experiments are presented.
to raptor and markedly decreased phosphorylation of PHAS-I by mTOR in vitro (29,(31)(32)(33). The fact that PHAS-I and PHAS-II each contain both motifs may account for similarities in their regulation by insulin. As previously proposed by Wang et al. (26), a lack of the RAIP motif may explain the relative insensitivity of PHAS-III to insulin.
The present finding that Ser-64 phosphorylation is dispensable for insulin-dependent control of the PHAS-I⅐eIF4E complex was surprising in view of previous observations. Ser-64 was identified as the site undergoing the largest increase in phosphorylation in response to insulin even before PHAS-I was identified as an eIF4E-binding protein (12). Initial studies demonstrated that selectively phosphorylating Ser-64 abolished high affinity binding of PHAS-I to eIF4E (3,10). More recent measurements of binding affinity of purified proteins by surface plasmon resonance indicate that phosphorylation of this site decreases the affinity for eIF4E by ϳ100-fold (14). Changes in affinity due to phosphorylation of other sites individually were relatively small (11,14). Finally, there is the phosphorylation hierarchy in which the three Thr-Pro sites in PHAS-I must be phosphorylated before Ser-64 can be phosphorylated (9,13). These observations contributed to an attractive model in which progressive phosphorylation of the three Thr-Pro sites culminates in the phosphorylation of Ser-64 and dissociation of the PHAS-I⅐eIF4E complex (1). The present findings indicate that this model is incorrect in the sense that Ser-64 is dispensable for insulin-stimulated dissociation of the complex (Fig. 2).
Although phosphorylating the Thr-Pro sites in PHAS-I indi- vidually had small effects on the affinity of PHAS-I for eIF4E relative to that of phosphorylating Ser-64 (11,14), it is possible that phosphorylating a combination of the Thr-Pro sites in PHAS-I would produce more pronounced effects. Consequently, the present conclusion that phosphorylation of the Thr-Pro sites is sufficient for insulin-stimulated dissociation of the PHAS-I⅐eIF4E complex is not inconsistent with previous studies (11,14), which did not address the effect of phosphorylating multiple Thr-Pro sites. We cannot exclude the possibility that Ser-64 phosphorylation contributes to the control of the association state of PHAS-I and eIF4E in response to stimuli other than insulin or in manner too subtle to allow detection with the current assays. However, Ser-64 phosphorylation might control other functions of PHAS-I such as PHAS-I binding to raptor. Mutating the five (Ser/Thr)-Pro sites markedly increased the binding of PHAS-I to raptor, suggesting that phosphorylation of PHAS-I promotes dissociation of the PHAS-I raptor complex (33).
Diggle et al. (34) were the first to show that PHAS-I was phosphorylated by protein kinase CK2 (34). Fadden et al. (15) identified Ser-111 as the major site of phosphorylation by this kinase in vitro but questioned the role of this site because its phosphorylation was not detected when PHAS-I was immunoprecipitated from 32 P-labeled rat adipocytes. A possible explanation for the failure to detect Ser-111 phosphorylation in cells was provided by Denton's group, who observed significant phosphorylation of the site only in the PHAS-I that was bound to eIF4E (8). This finding led to the proposal that Ser-111 serves as a priming site whose phosphorylation is necessary for the phosphorylation of other sites in PHAS-I and for dissociation of the PHAS-I⅐eIF4E complex. Such a role gained acceptance as a result of a report that mutation of Ser-111 to Ala decreased the phosphorylation of the other sites in PHAS-I overexpressed 293T cells (18). The present findings clearly do not support this model. We observe no difference in the electrophoretic mobilities of wild type PHAS-I and Ala-111 PHAS-I, indicating that mutation of Ser-111 does not affect the phosphorylation of sites involved in the mobility shift (Fig. 3). Results of multiple experiments with phosphospecific antibodies demonstrate that phosphorylation of Thr-36/45, Thr-69, and Ser-64 were not influenced by mutating Ser-111 to Ala (Fig. 4).
FIG. 6. Effects of insulin, rapamycin, and amino acids on the association of PHAS-II and eIF4E. HEK293 cells were transfected with vectors alone (pCMV-Tag 2B and pCMV-Tag 3A) or with pCMV-Tag 2B eIF4E plus pCMV-Tag3A PHAS-II . The cells were incubated as described in the legend to Fig. 1 before extracts were prepared and Myc-tagged PHAS-II was immunoprecipitated with the 9E10 antibody. A, immunoblots were prepared with antibodies to PHAS-II and eIF4E. B, the amounts of total eIF4E (FLAG-tagged plus endogenous) coimmunoprecipitating with Myc-PHAS-II were estimated from optical density scans of eIF4E immunoblots. Results are expressed relative to control and are mean values Ϯ S.E. from five experiments. C, eIF4E was partially purified by using m 7 GTP-Sepharose. Relative amounts of PHAS-II copurifying with eIF4E are expressed relative to control and are mean values Ϯ S.E. from four experiments. Similar results were very recently reported by Wang et al. (26). These findings support the conclusion that Ser-111 phosphorylation is not required for the phosphorylation of other sites in PHAS-I. Yang and Kastan (18) proposed that phosphorylation of Ser-111 by ATM mediates insulin-stimulated dissociation of the PHAS-I⅐eIF4E complex. This hypothesis was both provocative and exciting, not only because it linked ATM-mediated checkpoint control, insulin action, and cap-dependent protein synthesis, but also because it appeared to identify ATM as an important kinase in insulin action. Although our studies have not directly addressed the role of ATM, it is clear that phosphorylation of Ser-111 is not essential in the control of PHAS-I because mutating Ser-111 to Ala did not prevent the dissociation of the PHAS-I⅐eIF4E complex in response to insulin (Fig.  5). The finding that insulin stimulated the dissociation of eIF4E from PHAS-II (Fig. 6), which lacks the Ser-111 site (2), provides confirmation that phosphorylation of the Ser-111 site is not essential for dissociation of PHAS⅐eIF4E complexes.