Effects of Protein Phosphorylation on Ubiquitination and Stability of the Translational Inhibitor Protein 4E-BP1

Home , Dephosphorylation, Protein phosphorylation

Oncogene (2008) 27, 811–822 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE Effects of protein phosphorylation on ubiquitination and stability of the translational inhibitor protein 4E-BP1

A Elia, C Constantinou1 and MJ Clemens

Translational Control Group, Division of Basic Medical Sciences, Centre for Molecular and Metabolic Signalling, St George’s, University of London, London, UK

The availability of the eukaryotic polypeptide chain structure by polypeptide chain initiation factor eIF4E initiation factor 4E (eIF4E) for protein synthesis is (Mamane et al., 2004; von der Haar et al., 2004). The regulated by the 4E-binding proteins (4E-BPs), which availability of the latter for formation of the active act as inhibitors of cap-dependent mRNA translation. The eIF4F complex is regulated by the 4E-binding proteins ability of the 4E-BPs to sequester eIF4E is regulated by (4E-BP1 and 4E-BP2) (Clemens, 2001; Mamane et al., reversible phosphorylation at multiple sites. We show here 2006). The most thoroughly studied of these is 4E-BP1, that, in addition, 4E-BP1 is a substrate for polyubiquiti- which in its hypophosphorylated state, binds and nation and that some forms of 4E-BP1 are simultaneously sequesters eIF4E. Phosphorylation of 4E-BP1, at multi- polyubiquitinated and phosphorylated. In Jurkat cells ple sites, decreases the affinity of this protein for eIF4E inhibition of proteasomal activity by MG132 enhances the and thus releases the cap-binding protein for eIF4F level of hypophosphorylated, unmodified 4E-BP1 but only complex formation. The phosphorylation of 4E-BP1 is modestly increases the accumulation of high-molecular- influenced by a wide range of extracellular stimuli. In weight, phosphorylated forms of 4E-BP1. In contrast, general, conditions that promote cell growth enhance inhibition of protein phosphatase activity with calyculin A the phosphorylation of 4E-BP1 and are associated with reduces the level of unmodified 4E-BP1 but strongly increased rates of overall protein synthesis; conversely, enhances the amount of phosphorylated, high-molecular- growth inhibitory conditions and physiological stresses weight 4E-BP1. Turnover measurements in the presence result in dephosphorylation of the protein and cause of cycloheximide show that, whereas 4E-BP1 is normally downregulation of protein synthesis (Mamane et al., a very stable protein, calyculin A decreases the apparent 2006). The protein kinase mammalian target of rapa- half-life of the normal-sized protein. Affinity chromato- mycin (mTOR) plays a crucial role in the phosphoryla- graphy on m7GTP-Sepharose indicates that the larger tion of 4E-BP1, although other protein kinases have also forms of 4E-BP1 bind very poorly to eIF4E. We suggest been implicated (Herbert et al., 2002; Gingras et al., that the phosphorylation of 4E-BP1 may play a dual role 2004; Tee and Blenis, 2005). There is evidence that in the regulation of protein synthesis, both reducing the 4E-BP1 associates with protein phosphatase 2A (Peter- affinity of 4E-BP1 for eIF4E and promoting the conver- son et al., 1999), and it is likely that the depho- sion of 4E-BP1 to alternative, polyubiquitinated forms. sphorylation of 4E-BP1 is catalysed by this enzyme. Oncogene (2008) 27, 811–822; doi:10.1038/sj.onc.1210678; The availability of eIF4E has differential effects on published online 23 July 2007 the translation of different mRNAs, with those containing complex, structured 50-untranslated regions being Keywords: calyculin A; eIF4E; protein degradation; protein particularly sensitive to changes in the level of the factor phosphatases; protein synthesis (Graff and Zimmer, 2003). The initiation factor also exerts important effects on the nucleo-cytoplasmic transport of a number of mRNAs (Strudwick and Borden, 2002). Overexpression of eIF4E is associated Introduction with malignant transformation (De Benedetti and Graff, 2004), probably because the factor promotes the The translation of most mRNAs in eukaryotic cells synthesis of antiapoptotic proteins (Clemens, 2004; requires recognition of the 50 m7GTP-containing cap Mamane et al., 2004), and many naturally occurring tumours contain elevated levels of eIF4E. Conversely, enhanced expression of 4E-BP1 can reverse the trans- Correspondence: Professor MJ Clemens, Translational Control Group, Division of Basic Medical Sciences, Centre for Molecular formed phenotype and increase the susceptibility of cells and Metabolic Signalling, St George’s, University of London, to apoptosis (Li et al., 2002; Proud, 2005). In naturally Cranmer Terrace, London SW17 0RE, UK. occurring tumours, the level of phosphorylation of E-mail: [email protected] 4E-BP1 correlates with tumour stage and prognosis 1Current address: Yasoo Health Ltd, 1 Poseidon Street, PO Box 25193, 1307 Nicosia, Cyprus. (Castellvi et al., 2006). Received 22 February 2007; revised 23 May 2007; accepted 13 June 2007; Whereas a large body of work has been devoted to the published online 23 July 2007 phosphorylation of 4E-BP1, very few studies have Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 812 examined regulation of the stability of the protein. 1995; Gingras et al., 1996; Mendez et al., 1996), mouse Following fertilization of sea urchin eggs, there is a rapid and human 4E-BP1 were detected as multiple bands on loss of 4E-BP which occurs in parallel with dramati- gels in the vicinity of 20 kDa. These correspond to cally enhanced rates of protein synthesis (Cormier differentially phosphorylated forms, with hyperpho- et al., 2001; Salau¨ n et al., 2003). This loss is inhibited sphorylated (g) 4E-BP1 migrating more slowly than by rapamycin, suggesting a role for protein phosphory- the less phosphorylated (b and a) forms. In addition, lation in the degradation of 4E-BP. In a mammalian analysis of extracts from mouse embryonic fibroblasts cell system dephosphorylation of 4E-BP1 induced by showed that regions of the blots corresponding to the tumour suppressor protein p53 is accompanied higher-molecular-weight proteins contain several other by an increase in the level of the 4E-BP1, in parallel with bands that are recognized by antibodies against total inhibition of protein synthesis (Tilleray et al., 2006). 4E-BP1 (Figure 1a). At least some of these forms of the However, the mechanisms regulating 4E-BP1 levels have protein are ubiquitinated, as revealed by immunopreci- not so far been extensively investigated. pitation with anti-4E-BP1 followed by immunoblotting Many proteins of regulatory significance are targeted with anti-ubiquitin. As shown in Figure 1b (left panel), a for degradation by the 26S proteasome as a result of band that migrates at around 50 kDa was identifiable their polyubiquitination (Fang and Weissman, 2004). by this approach. When the converse experiment was Moreover, the susceptibility of such proteins to be performed, immunoprecipitating a cell extract with anti- processed by this pathway is often controlled by protein ubiquitin-coated agarose beads and blotting with anti- phosphorylation. In addition to the now classical cases 4E-BP1, a similar sized band was seen (Figure 1b, right of the inhibitor of nuclear factor kB(IkB) family (Karin panel). The mobility of this protein suggests that it contains and Ben-Neriah, 2000; Chen, 2005; Krappmann and three or four ubiquitin moieties. The co-precipitation Scheidereit, 2005) and cyclins (Lin et al., 2006a), other reaction was specific since no anti-4E-BP1 reactive material examples of proteins whose (poly)ubiquitination is was observed when unmodified agarose beads were used stimulated by their phosphorylation include the type I (Figure 1b, right panel, lane 2) and no unmodified interferon receptor IFNAR1 (Kumar et al., 2004), the (20 kDa) 4E-BP1 was detected after anti-ubiquitin transcription factor STAT1 (Kim and Maniatis, 1996), precipitation. the antiapoptotic protein Bcl-2 (Lin et al., 2006b) and b- catenin (Aberle et al., 1997; Orford et al., 1997). On the

other hand, phosphorylation can inhibit the ubiquitina- kDa M tion of other proteins such as the proto-oncogenes c-jun 100 80 (Musti et al., 1997; Fuchs et al., 1998), c-fos (Okazaki 60 Higher molecular and Sagata, 1995) and c-mos (Nishizawa et al., 1993). 50 40 Recent studies have shown that initiation factor forms 30 eIF4E can undergo ubiquitination and proteasome- γ β mediated turnover (Othumpangat et al., 2005a; Murata 20 α and Shimotohno, 2006). In contrast, although there have been indications that 4E-BP1 is also degraded by 10 the proteasome (Walsh and Mohr, 2004; Walsh et al., 1 MkDa M 2 3 2005; Wan et al., 2005), no information has been 60 50 polyUb-4E-BP1 published on the possible ubiquitination of this protein. polyUb-4E-BP1 40 In this paper, we present evidence from a variety of cell 30 lines indicating that human and mouse 4E-BP1 can be 20 polyubiquitinated, giving rise to high-molecular-weight 10 Anti-4E-BP1 Anti-UbIP forms. Our data suggest that this process is controlled IPAnti-Ubblot Anti-4E-BP1 blot by phosphorylation of the protein. Moreover, we show Figure 1 Evidence for polyubiquitination of 4E-BP1. (a)A that enhancing the phosphorylation of 4E-BP1 both cytoplasmic extract from murine embryonic fibroblasts was decreases the level of the normal-sized protein and subjected to gel electrophoresis followed by immunoblotting for converts the latter into high-molecular-weight forms total 4E-BP1. Molecular weight markers were separated in a that are unable to associate with eIF4E. The implica- parallel lane (M) and their sizes (in kDa) are indicated. The regions of the blot corresponding to the a, b and g forms of 4E-BP1 and to tions for the regulation of protein synthesis and cell higher-molecular-weight cross-reacting bands are shown. The transformation by eIF4E and 4E-BP1 are discussed. experiment was repeated several times and a typical blot is shown. (b) A cytoplasmic extract from fibroblasts was subjected to immunoprecipitation with anti-4E-BP1, anti-ubiquitin agarose beads or unmodified agarose beads. Left panel: the anti-4E-BP1 immunoprecipitate was analysed by gel electrophoresis followed by Results immunoblotting with anti-ubiquitin (lane 1). Right panel: the unmodified agarose beads (lane 2) and anti-ubiquitin agarose beads Identification of high-molecular-weight forms of 4E-BP1 (lane 3) were incubated in 2 Â SDS sample buffer and the eluted We have used immunoblotting with several antibodies proteins analysed by gel electrophoresis followed by immunoblotting with anti-4E-BP1. The sizes of molecular weight markers (in against total 4E-BP1 to characterize the size distribution kDa) are indicated. The regions of the blots corresponding to high- of the protein. In agreement with numerous previous molecular-weight (polyubiquitinated) forms of 4E-BP1 are shown. studies (see for example, Lin et al., 1994; Graves et al., 4E-BP1, 4E-binding protein 1; SDS, sodium dodecyl sulphate.

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 813 Relationship between phosphorylation and To test directly whether the high-molecular-weight ubiquitination of 4E-BP1 forms of 4E-BP1 that accumulate in the presence of Since 4E-BP1 is modified by phosphorylation and can calyculin A (7MG132) are both phosphorylated and also become polyubiquitinated, we have investigated ubiquitinated, extracts from Jurkat cells were immuno- whether there is any interaction between the two events, precipitated with anti-ubiquitin and then immuno- and whether the same protein molecules can undergo blotted for phospho-Thr37/46. Figure 3 shows that the both types of modification. Cells with relatively low immunoprecipitation brought down high-molecular- basal levels of 4E-BP1 phosphorylation (murine ery- weight proteins of ca. 50 and 80 kDa that cross-reacted throleukaemia (MEL) cells in which a temperature- with anti-phospho-Thr37/46. Moreover, calyculin sensitive form of p53 had been activated (Constantinou A þ MG132 strongly enhanced the phospho-Thr37/46 and Clemens, 2005)) were treated with the protein signal, in both total cytoplasmic extracts and the phosphatase inhibitor calyculin A. Extracts were then immunoprecipitated material, and an even larger form examined for the appearance of phosphorylated forms (ca. 200 kDa) was observed under these conditions. of the protein. Immunoblotting with phospho-specific These results confirm that at least some forms of high- antibodies showed that calyculin A had a small molecular-weight 4E-BP1 are ubiquitinated and phos- stimulatory effect on the level of phosphorylation of phorylated on the same molecules. normal-sized mouse 4E-BP1, particularly on Ser64 (Figure 2a, compare lanes 1 and 3). More dramatically, Relationship between ubiquitination and association the phosphatase inhibitor markedly enhanced the of 4E-BP1 with eIF4E appearance of phosphorylated high-molecular-weight We have shown that the high-molecular-weight forms of forms of the protein. In this case, increased phosphor- 4E-BP1 that appear in response to calyculin A in Jurkat ylation was observed at all sites examined, with a cells are highly phosphorylated on Thr37/46. There is also particularly strong effect on Thr36/45. Several bands, with phosphorylation on Ser65 but relatively little on Thr70 approximate molecular masses in the range of 50– (Figure 2b). Since the latter two sites have important 200 kDa, were observed. Similar results were observed roles in regulating the association of 4E-BP1 with eIF4E using human Jurkat T-lymphoma cells (Figure 2b). it was unclear whether ubiquitinated forms of 4E-BP1 These data suggest that 4E-BP1 may be both phos- would be present in complexes with the initiation factor. phorylated and polyubiquitinated simultaneously. In To determine this, eIF4E and its associated proteins contrast to the results with phospho-specific antibodies, were purified by m7GTP-Sepharose affinity chromato- calyculin A had very little effect on the appearance of graphy from extracts of cells incubated with or without additional forms of 4E-BP1 when an antibody that calyculin A for different periods of time. Figure 4a recognizes the total protein was used, suggesting that shows that, after 1 h of calyculin A treatment, almost only a relatively small fraction of 4E-BP1 becomes none of the high-molecular-weight 4E-BP1 that was polyubiquitinated in the presence of the phosphatase phosphorylated on Thr37/46 could be found in association inhibitor. However the 4E-BP1 that was modified in this with eIF4E. This was in contrast to the behaviour of manner was heavily phosphorylated, particularly on normal-sized 4E-BP1 where the very small amount still Thr36/45 in MEL cells (Thr37/46 in Jurkat cells). associated with eIF4E in calyculin A-treated cells did Because ubiquitination can tag proteins for degrada- include material phosphorylated on Thr37/46 (but not on tion via the proteasome pathway, we investigated Ser64 or Thr69—data not shown). Similarly, after a whether inhibition of proteasome activity had any longer period of incubation with calyculin A only a trace influence on the levels of the normal-sized and high- amount of higher-molecular-weight 4E-BP1 was bound molecular-weight forms of 4E-BP1. Exposure of both to eIF4E (Figure 4b). Analysis of eIF4E-associated MEL cells and Jurkat cells to the proteasome inhibitor proteins also failed to detect any ubiquitinated species MG132 had two notable effects. First MG132 caused (data not shown). These results suggest that the substantial dephosphorylation of the normal-sized ubiquitination of 4E-BP1 (but not the phosphorylation protein, especially on Ser64 (Ser65 in Jurkat cells) and of Thr37/46 per se) reduces the affinity of the protein enhanced the level of the hypophosphorylated (a) form, for eIF4E. particularly in Jurkat cells (Figure 2b). Secondly, MG132 modestly stimulated the appearance of high- molecular-weight forms of 4E-BP1 that were phos- Phosphorylation and the stability of 4E-BP1 phorylated on Thr36/45 (Thr37/46) (Figures 2a–c, compare Although calyculin A increased the phosphorylation of lanes 1 and 2). The latter effect may reflect inhibition of the high-molecular-weight bands (especially on Thr37/46), proteasome-mediated degradation of phosphorylated, it reduced the total level of normal-sized 4E-BP1 in ubiquitinated forms of 4E-BP1. However, our data cytoplasmic extracts, especially in the case of Jurkat cells suggest that the phosphorylation of 4E-BP1 drives the (Figure 2b, top panel and Figure 4). Analysis of ubiquitination process; thus the limited ability of MG132 cytoplasmic and nuclear fractions from the cells did to cause accumulation of ubiquitinated 4E-BP1, relative not indicate any change in the subcellular distribution of to the effect of the inhibitor on total ubiquitinated the protein in response to calyculin A (data not shown). proteins (Figure 2c, middle panel) may be due to the To test the effect of calyculin A on the half-life of separate effect of MG132 in inhibiting the phosphorylation normal-sized 4E-BP1 levels of the protein were mea- of normal-sized 4E-BP1. sured after different times in the presence of the protein

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 814 synthesis inhibitor cycloheximide. The top panel in protein disappeared with an apparent half-life of about Figure 5a shows that under the control conditions, 4E- 9 h (Figure 5a, middle panel). This suggests that the BP1 was a relatively stable protein, with a half-life well protein phosphatase inhibitor either enhances the in excess of 16h. A previous study in adipocytes has also conversion of 4E-BP1 to larger forms or promotes the indicated that 4E-BP1 is stable (Lin et al., 1995). In complete destruction of the protein. The latter effect contrast, in the presence of calyculin A the normal-sized would be consistent with enhanced ubiquitination and

Calyculin A --++ Calyculin A --++ MG132 -+-+ MG132 -+-+ kDa kDa 50 - - 60 40 - 40 - Total 4E-BP1 γ Total 4E-BP1 30 - γ - β 20 α 20 - β α

200 - 200 - 100 - Higher molecular 100 - Higher molecular 60 - weight forms weight forms (P)Ser64 (P)Ser65 60 - - 40 40 - - γ 20 20 - γ

200 - 200 - - 100 100 - Higher molecular 60 - Higher molecular weight forms weight forms 70 60 - 69 (P)Thr (P)Thr - 40 40 - γ 20 - γ 20 - β

- 200 - 200 100 - 100 - Higher molecular Higher molecular weight forms 37/46 60 - weight forms 60 - (P)Thr - 40 - 40 γ 36/45 γ (P)Thr β 20 - 20 - β α α

α α-tubulin -tubulin 142 3 142 3

Calyculin A --++ MG132 -+-+ kDa

200 - 140 - 100 - Higher molecular 80 - weight forms 4E-BP1 60 - 37/46 50 - (P)Thr 40 - 30 - γ 20 - β α

200 - 140 - 100 - 80 - Ubiquitin 60 - 50 - 40 - 30 -

20 -

α-tubulin

1 2 3 4

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 815 subsequent degradation of phosphorylated 4E-BP1. The 4E-BP1, and MG132 did not prevent this. The proteasome extracts were also probed with the antibody against inhibitor did slightly increase the calyculin-induced accu- phospho-Thr37/46. As can be seen in Figure 5b, this mulation of the larger form(s), suggesting a possible role analysis revealed that the phosphorylated high-molecu- for the proteasome in the degradation of high-molecular- lar-weight forms of 4E-BP1 appeared within 1–2 h of weight, polyubiquitinated 4E-BP1. However, the latter exposure of cells to the phosphatase inhibitor and were forms were clearly relatively stable, indicating that maintained throughout the time course with cyclohex- ubiquitination per se does not lead to rapid turnover. It imide, declining slightly by 16h. However, these data do is possible that other components of the cell’s proteolytic not provide accurate information about the half-lives machinery are involved in the degradation of polyubiqui- of these phosphorylated forms of the protein since a tinated 4E-BP1. For example, lysosomes have been reservoir of phosphorylated normal-sized 4E-BP1 demonstrated to degrade a variety of polyubiquitinated remained, from which further high-molecular-weight proteins (Hicke, 2001; Kumar et al., 2004). However, molecules could have been derived. treatment of Jurkat cells with the lysosomal inhibitor To investigate a possible role for the proteasome in the chloroquine did not prevent the calyculin A-induced loss of degradation of ubiquitinated 4E-BP1 the effects of long- normal-sized 4E-BP1 or cause any further accumulation of term treatment with MG132 (7calyculin A) on the the larger forms (data not shown). normal-sized and higher-molecular-weight forms of Calyculin A inhibits the dephosphorylation of a wide 4E-BP1 were also examined in the presence of cyclohex- range of protein phosphatase substrates (Ishihara et al., imide. Since extended exposure to MG132 can induce 1989) and it was possible that it could non-speciﬁcally apoptosis (Lin et al., 1998) in this experiment the caspase alter the activity of a ubiquitin ligase and/or affect inhibitor benzyloxycarbonyl-Val-Ala-Asp-ﬂuoromethylke- components of the protein degradation machinery itself. tone (z.VAD-FMK) was added to the cells to eliminate To determine whether the effects of calyculin A on any possible role of caspase activation in the degradation 4E-BP1 require the mTOR-mediated phosphorylation of 4E-BP1. Figure 5c shows that, after 16h, calyculin of the protein, we therefore examined the fate of 4E-BP1 treatment strongly reduced the level of normal-sized in the additional presence of rapamycin. This agent is a

[Calyculin A + MG132] -+ - + kDa kDa 200 Higher molecular 100 100 weight forms 80 60 60 (P)Thr 37/46 50 50 40 40 30 γ 30 20 β α 20

Cell extracts Anti-Ub IP Figure 3 Evidence that 4E-BP1 can be simultaneously phosphorylated and ubiquitinated. Jurkat cells were incubated with or without MG132 for 16h. Where indicated, calyculin A was then added for the last hour. Cytoplasmic extracts were prepared and samples were immunoprecipitated with anti-ubiquitin agarose beads. The extracts (left panel) and immunoprecipitates (right panel) were then analysed by gel electrophoresis and immunoblotting for phosphorylated 4E-BP1 (Thr37/46). The sizes of molecular weight markers (in kDa) are indicated. The regions of the blots corresponding to the a, b and g forms of 4E-BP1 and to higher-molecular-weight forms of the protein are shown. (Note that the lowest band in the left panel is a smaller fragment of 4E-BP1 that is often observed in Jurkat cell extracts.) 4E-BP1, 4E-binding protein 1.

Figure 2 Effects of protein phosphatase and proteasome inhibitors on 4E-BP1. (a) MEL cells expressing a temperature-sensitive form of p53 were incubated at 321C for 16h to invoke p53-dependent dephosphorylation of 4E-BP1 (Constantinou and Clemens, 2005) and treated with calyculin A and/or MG132 where indicated. Cytoplasmic extracts were subjected to gel electrophoresis followed by immunoblotting for total 4E-BP1, phosphorylated 4E-BP1 (Ser64, Thr69 and Thr36/45) and a-tubulin as indicated. (b) Jurkat cells were incubated with or without MG132 for 16h. Where indicated, calyculin A was then added for the last hour. Cytoplasmic extracts were prepared and analysed as in (a). (c) Jurkat cells were incubated with or without MG132 and/or calyculin A for 2 h. Cytoplasmic extracts were prepared and analysed by immunoblotting for phosphorylated 4E-BP1 (Thr37/46), total ubiquitinated proteins and a-tubulin as indicated. In all experiments, molecular weight markers were separated in parallel lanes and their sizes (in kDa) are indicated. The regions of the blots corresponding to the a, b and g forms of normal-sized 4E-BP1 and to higher-molecular-weight forms of the protein are shown. The experiments were repeated several times and typical blots are shown. For the phospho-Thr37/46 blots in (b) and (c) different exposures have been used for lanes 1 and 2, and 3 and 4, respectively to optimize visualization of the bands. Note that the bands at ca. 30 and 50 kDa in the top panel of (b) are non-speciﬁc Jurkat cell proteins that cross-react with the antibody used and are not ubiquitinated forms of 4E-BP1. 4E-BP1, 4E-binding protein 1.

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 816 Calyculin A -+ -+ -+ -+ kDa kDa kDa kDa Higher molecular 50 - weight forms 50 - 40 - 40 - 50 - 50 - 40 - 30 - 40 - 30 - 30 - 30 - 20 - 20 - 20 - 20 - -+

Total 4E-BP1 (P)Thr37/46 Total 4E-BP1 (P)Thr37/46 eIF4E

Cytoplasmic extracts m7GTP fractions

Calyculin A - + -+ -+ -+ kDa kDa kDa Higher molecular weight forms 50 - 50 - 50 - 50 - 40 - 40 - 40 - 40 - 30 - 30 - 30 - 30 - 20 - 20 - 20 - 20 - -+

37/46 Total 4E-BP1 (P)Thr Total 4E-BP1 (P)Thr37/46 eIF4E

Cytoplasmic extracts m7GTP fractions Figure 4 Analysis of the association of modified forms of 4E-BP1 with eIF4E. Jurkat cells were incubated with or without MG132 for 16h. Where indicated, calyculin A was added for the last hour ( a) or was present throughout the incubation (b). Cytoplasmic extracts were prepared and eIF4E and its associated proteins were isolated by m7GTP-Sepharose affinity purification. Samples of the affinity purified fractions were analysed for total and phosphorylated 4E-BP1 (Thr37/46) by gel electrophoresis and immunoblotting. The affinity purified fractions were also blotted for eIF4E as a control for equal protein recovery. 4E-BP1, 4E-binding protein 1.

direct inhibitor of mTOR and partially lowers the level degradation of 4E-BP1 in cells infected by herpes of phosphorylation of 4E-BP1 (Gingras et al., 2004). simplex virus-1. Furthermore, their data suggested that Figure 6(top panel) shows that rapamycin partially degradation was sensitive to the state of phosphoryla- prevented the calyculin A-induced decrease in the level tion of 4E-BP1. However, it is not yet clear whether 4E- of normal-sized 4E-BP1 in Jurkat cells. Although the BP1 is degraded exclusively by the proteasome since mTOR inhibitor was not fully able to reverse this effect interpretation of results from experiments using MG132 of calyculin A, it also had only a modest effect on the to inhibit proteasomal activity is complicated by the fact phosphorylation of 4E-BP1 (mostly on Ser65 and Thr70), that this agent also causes the dephosphorylation of one as well as on the calyculin-induced appearance of the or more sites on 4E-BP1 (Figures 2a and b). This could phosphorylated high-molecular-weight forms. Never- be an indirect effect reflecting inhibition of a protein theless, our findings suggest that the disappearance of kinase or stabilization of a protein phosphatase by normal-sized 4E-BP1 following calyculin A treatment is MG132 (Torres et al., 2003) and could potentially likely to be a direct consequence of increased phosphory- inhibit the ubiquitination process that is driven by lation, rather than a non-specific change in the activity phosphorylation. In addition, as indicated in Figure 7, of the protein ubiquitination or degradation machinery. there may be a role for (poly)ubiquitination of 4E-BP1 in controlling properties of this factor besides stability, as is the case for several other proteins of regulatory significance (Aguilar and Wendland, 2003; Chen, 2005). Discussion In this connection it is of interest that the high- molecular-weight forms of 4E-BP1 bind very poorly to Our data suggest a model in which the phosphorylation eIF4E (Figure 4). of 4E-BP1 regulates the ubiquitination of this protein. Since inhibition of protein phosphatase activity with As shown in Figure 7, this may promote the turnover of calyculin A leads to the accumulation of phosphorylated a fraction of 4E-BP1, mediated by the proteasome and/ high-molecular-weight forms of 4E-BP1, an important or other proteolytic mechanisms, but may have other question is why very little of the phosphorylated protein consequences as well. A previous study (Walsh and occurs in these forms under normal circumstances. One Mohr, 2004) proposed a role for the proteasome in the possibility is that hyperphosphorylation of 4E-BP1

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 817 Time (h) 0 1 2 4 8 16 Time (h) 0 1 2 4 8 16 kDa kDa γ + CHX 60 - + CHX 20 - β 50 - α 40 -

30 - + CHX γ γ 20 - β 20 - β + cal A α α

Total 4E-BP1 + CHX 100 Higher molecular + cal A 60 - weight forms 50 - 40 - 50 Control forms) forms)

γ 30 - β

and 20 - + calyculin A α β

, α, β α Level of 4E-BP1 Level ( (% of initial value) 37/46 10 (P)Thr 0 5 10 15 20 Time with cycloheximide (h)

Calyculin A - + - + MG132 - - + + kDa

100 - Higher molecular 80 - weight forms 60 - 50 - Total 4E-BP1 40 - 30 -

γ 20 - β α

α-tubulin

Figure 5 Effects of calyculin A and MG132 on the stability of 4E-BP1 and conversion to high-molecular-weight forms. (a) Jurkat cells were incubated with or without calyculin A for 1 h and cycloheximide was then added for the times indicated. Upper panel: cytoplasmic extracts were analysed by gel electrophoresis and immunoblotting using antibody against total 4E-BP1. The sizes of molecular weight markers (in kDa) are indicated. The regions of the blots corresponding to the a, b and g forms of 4E-BP1 are shown. (Note that the band below the a form of 4E-BP1 is a smaller fragment that is often observed in Jurkat cell extracts.) Bottom panel: the combined relative intensities of the a, b and g forms of 4E-BP1 (determined by quantitative densitometry) are plotted on a log scale as a function of time with cycloheximide. The intensity corresponding to 50% of the initial value is indicated by the horizontal line. (b) The extracts analysed in (a) above were re-probed using antibody against phospho-Thr37/46.(c) Jurkat cells were pre-incubated with zVAD-FMK for 1 h to prevent subsequent apoptosis and then treated with or without calyculin A and/or MG132 in the presence of cycloheximide for 16h. Cytoplasmic extracts were analysed by gel electrophoresis and immunoblotting for total 4E-BP1 and a-tubulin. The a, b and g forms of normal-sized 4E-BP1 and higher-molecular-weight forms of the protein are indicated. 4E-BP1, 4E-binding protein 1.

induced by calyculin A enhances the accumulation of normal-sized 4E-BP1 (Figure 6) argues against less polyubiquitinated forms to an extent that exceeds the direct effects of the protein phosphatase inhibitor, for capacity of the protein degradation machinery to example on the enzymes involved in the ubiquitination rapidly degrade the protein. This would be consistent or degradation of 4E-BP1. However, the presence of a with the calyculin-induced decrease in the level of substantial amount of 4E-BP1 that is phosphorylated on normal sized 4E-BP1 (Figures 2 and 4) and the limited Thr37/46, Ser65 and Thr70 but which remains of normal ability of MG132 to increase further the levels of high- size in the presence of calyculin A (Figures 2, 4–6) molecular-weight phosphorylated 4E-BP1 (Figure 2). suggests that, although phosphorylation may be neces- Alternatively, an additional site on 4E-BP1, which is not sary, it is not sufﬁcient for polyubiquitination. It normally phosphorylated (or is phosphorylated in only remains to be determined whether the effects of a small fraction of the protein), may become more calyculin A are due to inhibition of the dephosphoryla- extensively phosphorylated in the presence of calyculin tion of 4E-BP1 itself or to an ability to enhance the A and act as a signal for polyubiquitination or other activity of mTOR (Carroll et al., 2006). modiﬁcations. The ability of rapamycin partially to Our data suggest that, under some circumstances, impair the effect of calyculin A in inducing the loss of stimuli that promote the phosphorylation of 4E-BP1

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 818 Calyculin A - + - + hypophosphorylated 4E-BP1 (Tilleray et al., 2006) and a Rapamycin - - + + similar outcome has been observed in cells treated with kDa the proapoptotic cytokine TRAIL (IW Jeffrey and MJ 60 - Clemens, unpublished data). The effects of genotoxic 50 - 40 - stress on 4E-BP1 are further discussed below. Total 4E-BP1 30 - γ There are many precedents in the literature for the 20 - β stimulation of protein ubiquitination and degradation α by phosphorylation (reviewed in Fuchs et al., 1998). However, there are also cases where phosphorylation 50 - and ubiquitination of a protein appear to be mutually 40 - Higher molecular weight forms exclusive. Of interest in the present context is the ﬁnding (P)Ser65 30 - that a ubiquitinated form of initiation factor eIF4E is 20 - never phosphorylated (Murata and Shimotohno, 2006). γ β This suggests that the ubiquitin-dependent turnover of eIF4E could be blocked by phosphorylation. Thus the phosphorylation of both eIF4E and 4E-BP1 that occurs in response to mitogenic and growth-promoting stimuli 60 - (P)Thr70 (Proud, 2002; Mamane et al., 2006) may have opposite 40 - consequences for the stability and levels of these two proteins. This would allow increased rates of protein 20 - γ β synthesis under growth-stimulatory conditions as a result of both enhanced stability of eIF4E and inactiva-

100 - Higher molecular tion or loss of its inhibitor. Conversely, physiological 60 - weight forms stresses that cause dephosphorylation of both eIF4E 37/46 50 - (P)Thr 40 - and 4E-BP1 may result in a decreased level of eIF4E and 30 - γ an increased level of 4E-BP1 due to opposite effects on 20 - β the turnover of these proteins. Consistent with this, α exposure of cells to ionizing radiation or DNA- damaging agents, which causes dephosphorylation of α-tubulin 4E-BP1 (Kumar et al., 2000; Tee and Proud, 2000), 1 2 3 4 increases the stability of the protein (Schneider et al., 2005; Le Bouffant et al., 2006). On the other hand, Figure 6 Effects of rapamycin on calyculin A-induced changes in physiological stresses such as heat shock or cadmium the level and phosphorylation of 4E-BP1. Jurkat cells were pre- incubated for 1 h in the presence or absence of rapamycin and/or chloride increase the ubiquitination and turnover of calyculin A. Cycloheximide was then added to the cells and the eIF4E (Othumpangat et al., 2005a; Murata and incubations were continued for a further 16h. Cell extracts were Shimotohno, 2006). However, DNA damage has no prepared and immunoblotted for total 4E-BP1, phosphorylated apparent effect on eIF4E turnover (Paglin et al., 2005). 65 70 37/46 4E-BP1 (Ser , Thr and Thr ) and a-tubulin as indicated. The a, Elevated levels of eIF4E result in inhibition of b and g forms of normal-sized 4E-BP1 and higher-molecular- weight forms of the protein are indicated. apoptosis (Clemens, 2004; Mamane et al., 2004), whereas overexpression of 4E-BP1 is proapoptotic (Li et al., 2002; Proud, 2005). It is therefore of interest that (for example, growth factors or other mitogenic agents) calyculin A is also antiapoptotic (Song and Lavin, may increase the ubiquitination and turnover of this 1993). Clearly, this protein phosphatase inhibitor is factor. Consistent with this, phosphorylation correlates likely to affect the phosphorylation of a wide range of with a dramatic loss of the protein following the additional target proteins that may have regulatory fertilization of sea urchin eggs (Cormier et al., 2001; effects on apoptosis (for example, IkB; Harhaj and Sun, Salau¨ n et al., 2003), and rapamycin has an inhibitory 1997). Nevertheless, the regulation of 4E-BP1 may effect on the loss of 4E-BP in this system. In contrast, in contribute to the inhibition of cell death by calyculin A. adipocytes insulin (which stimulates the phosphoryla- There are certain similarities between the regulation of tion of 4E-BP1) has been reported to increase the half- eIF4E function by 4E-BP1 and the control of nuclear life of the protein (Lin et al., 1995). However, this factor-kB(NF-kB) activity by IkB (Chen, 2005; Scheider- conclusion was based on pulse–chase studies of eit, 2006). Both IkB and 4E-BP1 are phosphorylated in [35S]methionine-labelled cells over a prolonged time response to extracellular signals, favouring the ubiquitina- period and it is possible that insulin may have directly tion and inactivation/degradation of these proteins. The inhibited protein degradation in these cells (Li et al., loss of IkB allows NF-kB to enter the nucleus and 2000; Fawcett et al., 2001a, b). stimulate the transcription of genes with antiapoptotic Conversely, physiological stresses that cause depho- activity (Graham and Gibson, 2005). Likewise, the sphorylation of 4E-BP1, such as amino-acid starvation, inactivation or loss of 4E-BP1 releases active eIF4E, DNA damage or the activation of p53 might be expected which stimulates post-transcriptionally the expression of to stabilize the protein. Recent data do indeed suggest antiapoptotic or oncogenic proteins (Rosenwald et al., that induction of p53 leads to the accumulation of 1995; Rousseau et al., 1996; Hoover et al., 1997; Carter

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 819

4E 4E (3) 4E 4E 4A 4E-BP1 rapamycin eIF4G (2) eIF4GeIF4G mTOR/ (4) (1) other kinases

S P (5) P T Translation 4E-BP1 4E-BP1 protein (High levels of eIF4E favour the translation of mRNAs phosphatase(s) encoding growth-promoting ++ (6) ubiquitination and anti-apoptotic proteins)

MG132 calyculin A P T S P 4E-BP1 (7) proteasome (8)

S P P T (9) DEGRADATION 4E-BP1 OTHER FUNCTIONS OF 4E-BP1 OF 4E-BP1?

Figure 7 Model for the regulation of 4E-BP1 by phosphorylation and polyubiquitination. In its hypophosphorylated state, 4E-BP1 binds to eIF4E (1) and sequesters the latter in an inactive complex. Phosphorylation of 4E-BP1 on multiple sites, by mTOR and other protein kinases (2), releases the eIF4E which is then available to interact with eIF4G (3) and initiate the translation of capped mRNAs (4). Phosphorylation is inhibited by rapamycin. High levels of active eIF4E preferentially favour the translation of mRNAs with extensive secondary structure, which encode growth-promoting and antiapoptotic proteins. Phosphorylated 4E-BP1 is dephosphorylated by protein phosphatase PP2A (5), thus allowing 4E-BP1 to become inhibitory for translation again. Dephosphorylation is inhibited by calyculin A. Our data suggest that phosphorylated 4E-BP1 can also be (poly)ubiquitinated (6). This may target some forms of the protein for degradation by the proteasome (7). Proteasome-mediated degradation is inhibited by MG132; this agent additionally causes dephosphorylation of normal-sized 4E-BP1 (presumably by an indirect mechanism) and this also stabilizes the protein. Some higher- molecular-weight forms of 4E-BP1 may subsequently undergo further modiﬁcations (8) to generate species that could have other functions in the cell (9). 4E-BP1, 4E-binding protein 1; mTOR, mammalian target of rapamycin.

et al., 1999; Nimmanapalli et al., 2003; Othumpangat Antibodies et al., 2005b). Since the overactivity or overexpression of Antibodies against total 4E-BP1 and a-tubulin were from either NF-kB or eIF4E inhibits apoptosis and promotes Santa Cruz Technology (Santa Cruz, CA, USA) and Sigma, respectively. Antibodies speciﬁc for phosphory- malignant transformation (Topisirovic et al., 2003; Clem- 64 36/45 69 ens, 2004; Ren et al., 2006), the eIF4E/4E-BP1 system and lation sites in 4E-BP1 (Ser , Thr and Thr in mouse; Ser65, Thr37/46 and Thr70 in human) were from Cell Signalling its regulatory mechanisms can be viewed as a translational Technology. Soluble anti-ubiquitin, anti-ubiquitin-agarose parallel to the NF-kB/IkBsystem. beads and unmodiﬁed agarose beads were from Calbiochem.

Materials and methods Preparation of cell extracts Cytoplasmic extracts were prepared by washing cells in Cell culture and treatments phosphate-buffered saline and re-suspending the pellets in cell Murine embryonic fibroblasts, MEL cells expressing a lysisbuffer(50mM 4-morpholinepropanesulphonic acid (MOPS), temperature-sensitive form of p53 (Johnson et al., 1993) and 50 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), Jurkat human T lymphoma cells were grown as described (Juo 2.5 mM ethylene glycol bis(b-aminoethylether)-N,N,N0,N0,- et al., 1998; Jeffrey et al., 2002; Constantinou and Clemens, tetraacetic acid (EGTA), 7 mM 2-mercaptoethanol, 40 mM 2005). Where specified, cells were treated with calyculin A (Cell b-glycerophosphate, 0.5% NP-40) containing the following Signaling Technology, Hitchin, Herts, UK) (10 nM), MG132 phosphatase and proteinase inhibitors: Complete Mini, EDTA- (Cell Signaling Technology) (30 mM), zVAD-FMK (Calbiochem, free protease inhibitor cocktail (Roche, Lewes, Sussex, UK), Nottingham, UK) (10 mM), rapamycin (Calbiochem) (30 nM) 2mM sodium vanadate, 1 mM microcystin, 1 mM phenylmethyl- or cycloheximide (Sigma, Poole, Dorset, UK) (20 mg/ml) for sulphonylfluoride and 2 mM benzamidine. After 10 min at 41C the times shown. Control cells received an equal volume of the cell lysates were centrifuged for 15 min at 12 000 g to pellet dimethylsulphoxide or water. the nuclei.

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 820 Immunoprecipitation and eIF4E affinity purification difluoride membranes. After fixation for 10 min with Immunoprecipitation was carried out in the presence of 20 mM 0.05% glutaraldehyde in phosphate-buffered saline the blots Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EGTA, 1 mM EDTA, were blocked with skimmed milk powder (5%, w/v) and 1% (v/v) Triton X-100, 2.5 mM Na pyrophosphate, 1 mM incubated with the appropriate primary antibodies. All b-glycerophosphate, 1 mM sodium vanadate and 1 mg/ml blots were developed with horseradish peroxidase-linked leupeptin, using either monoclonal anti-4E-BP1 and protein secondary antibodies using enhanced chemiluminescence A-agarose beads or anti-ubiquitin-agarose beads. Unmodified (Cell Signaling Technology). Levels of a-tubulin were deter- agarose beads were used as a control for non-specific mined as a loading control. Molecular weight markers were immunoprecipitation. After five washes in the above buffer a biotinylated protein ladder (10–200 kDa, from Cell Signaling bound proteins were eluted with 2 Â sodium dodecyl sulphate Technology) and were detected with horseradish peroxidase- (SDS) sample buffer. The proteins were then analysed by gel linked anti-biotin. Quantitative densitometry was performed electrophoresis and immunoblotting as described below. using Scion Image software (Scion Corporation, Frederick, Purification of eIF4E and associated proteins was per- Ma, USA). formed by affinity chromatography on m7GTP-Sepharose beads (GE Healthcare, Amersham, Bucks, UK) as described previously (Constantinou and Clemens, 2007). Acknowledgements

Immunoblotting We thank Dr Simon Morley (University of Sussex) for Cytoplasmic extracts were analysed by SDS gel electrophor- providing Jurkat cells and antibody to eIF4E. This work was esis, loading equal amounts of protein (10–20 mg per supported by the Association for International Cancer sample). The proteins were transferred to poly(vinylidene) Research and the Cancer Prevention Research Trust.

References

Aberle H, Bauer A, Stappert J, Kispert A, Kemler R. (1997). Fang S, Weissman AM. (2004). A ﬁeld guide to ubiquitylation. Beta-catenin is a target for the ubiquitin-proteasome path- Cell Mol Life Sci 61: 1546–1561. way. EMBO J 16: 3797–3804. Fawcett J, Hamel FG, Bennett RG, Vajo Z, Duckworth WC. Aguilar RC, Wendland B. (2003). Ubiquitin: not just for (2001a). Insulin and analogue effects on protein degradation proteasomes anymore. Curr Opin Cell Biol 15: 184–190. in different cell types. Dissociation between binding and Carroll M, Dyer J, Sossin WS. (2006). Serotonin increases activity. J Biol Chem 276: 11552–11558. phosphorylation of synaptic 4EBP through TOR, but Fawcett J, Hamel FG, Duckworth WC. (2001b). Character- eukaryotic initiation factor 4E levels do not limit somatic ization of the inhibition of protein degradation by insulin in cap-dependent translation in aplysia neurons. Mol Cell Biol L6cells. Arch Biochem Biophys 385: 357–363. 26: 8586–8598. Fuchs SY, Fried VA, Ronai Z. (1998). Stress-activated kinases Carter PS, Jarquin-Pardo M, De Benedetti A. (1999). regulate protein stability. Oncogene 17: 1483–1490. Differential expression of Myc1 and Myc2 isoforms in cells Gingras AC, Raught B, Sonenberg N. (2004). mTOR transformed by eIF4E: evidence for internal ribosome signaling to translation. Curr Top Microbiol Immunol 279: repositioning in the human c-myc 50UTR. Oncogene 18: 169–197. 4326–4335. Gingras AC, Svitkin Y, Belsham GJ, Pause A, Sonenberg N. Castellvi J, Garcia A, Rojo F, Ruiz-Marcellan C, Gil A, (1996). Activation of the translational suppressor 4E-BP1 Baselga J et al. (2006). Phosphorylated 4E binding protein 1: following infection with encephalomyocarditis virus and a hallmark of cell signaling that correlates with survival in poliovirus. Proc Natl Acad Sci USA 93: 5578–5583. ovarian cancer. Cancer 107: 1801–1811. Graff JR, Zimmer SG. (2003). Translational control and Chen ZJ. (2005). Ubiquitin signalling in the NF-kappaB metastatic progression: enhanced activity of the mRNA pathway. Nat Cell Biol 7: 758–765. cap-binding protein eIF-4E selectively enhances translation Clemens MJ. (2001). Translational regulation in cell stress and of metastasis-related mRNAs. Clin Exp Metastasis 20: apoptosis. Roles of the eIF4E binding proteins. J Cell Mol 265–273. Med 5: 221–239. Graham B, Gibson SB. (2005). The two faces of NF kappa B Clemens MJ. (2004). Targets and mechanisms for the in cell survival responses. Cell Cycle 4: 1342–1345. regulation of translation in malignant transformation. Graves LM, Bornfeldt KE, Argast GM, Krebs EG, Kong X, Oncogene 23: 3180–3188. Lin TA et al. (1995). cAMP- and rapamycin-sensitive Constantinou C, Clemens MJ. (2005). Regulation of the regulation of the association of eukaryotic initiation factor phosphorylation and integrity of protein synthesis initiation 4E and the translational regulator PHAS-I in aortic smooth factor eIF4GI and the translational repressor 4E-BP1 by muscle cells. Proc Natl Acad Sci USA 92: 7222–7226. p53. Oncogene 24: 4839–4850. Harhaj EW, Sun SC. (1997). The serine/threonine phosphatase Constantinou C, Clemens MJ. (2007). Regulation of transla- inhibitor calyculin A induces rapid degradation of I kappa B tion factors eIF4GI and 4E-BP1 during recovery of beta. Requirement of both the N- and C-terminal sequences. protein synthesis from inhibition by p53. Cell Death Differ J Biol Chem 272: 5409–5412. 14: 576–585. Herbert TP, Tee AR, Proud CG. (2002). The extracellular Cormier P, Pyronnet S, Morales J, Mulner-Lorillon O, signal-regulated kinase pathway regulates the phosphoryla- Sonenberg N, Belle´ R. (2001). eIF4E association with tion of 4E-BP1 at multiple sites. J Biol Chem 277: 4E-BP decreases rapidly following fertilization in sea urchin. 11591–11596. Dev Biol 232: 275–283. Hicke L. (2001). Protein regulation by monoubiquitin. Nat Rev De Benedetti A, Graff JR. (2004). eIF-4E expression and Mol Cell Biol 2: 195–201. its role in malignancies and metastases. Oncogene 23: Hoover DS, Wingett DG, Zhang J, Reeves R, Magnuson NS. 3189–3199. (1997). Pim-1 protein expression is regulated by its

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 821 50-untranslated region and translation initiation factor Lin T-A, Kong X, Saltiel AR, Blackshear PJ, Lawrence Jr JC. eIF-4E. Cell Growth Differ 8: 1371–1380. (1995). Control of PHAS-I by insulin in 3T3-L1 adipocytes. Ishihara H, Martin BL, Brautigan DL, Karaki H, Ozaki H, Synthesis, degradation, and phosphorylation by a rapamy- Kato Y et al. (1989). Calyculin A and okadaic acid: cin-sensitive and mitogen-activated protein kinase-indepen- inhibitors of protein phosphatase activity. Biochem Biophys dent pathway. J Biol Chem 270: 18531–18538. Res Commun 159: 871–877. Mamane Y, Petroulakis E, LeBacquer O, Sonenberg N. Jeffrey IW, Bushell M, Tilleray VJ, Morley S, Clemens MJ. (2006). mTOR, translation initiation and cancer. Oncogene (2002). Inhibition of protein synthesis in apoptosis: Differ- 25: 6416–6422. ential requirements by the tumor necrosis factor a family Mamane Y, Petroulakis E, Rong LW, Yoshida K, Ler LW, and a DNA-damaging agent for caspases and the double- Sonenberg N. (2004). eIF4E – from translation to transfor- stranded RNA-dependent protein kinase. Cancer Res 62: mation. Oncogene 23: 3172–3179. 2272–2280. Mendez R, Myers Jr MG, White MF, Rhoads RE. (1996). Johnson P, Chung S, Benchimol S. (1993). Growth suppres- Stimulation of protein synthesis, eukaryotic translation sion of Friend virus-transformed erythroleukemia cells by initiation factor 4Ephosphorylation, and PHAS-I phosphory- p53 protein is accompanied by hemoglobin production lation by insulin requires insulin receptor substrate and is sensitive to erythropoietin. Mol Cell Biol 13: 1 and phosphatidylinositol 3-kinase. Mol Cell Biol 16: 1456–1463. 2857–2864. Juo P, Kuo CJ, Yuan JY, Blenis J. (1998). Essential Murata T, Shimotohno K. (2006). Ubiquitination and requirement for caspase-8/FLICE in the initiation of the proteasome-dependent degradation of human eukaryo- Fas-induced apoptotic cascade. Curr Biol 8: 1001–1008. tic translation initiation factor 4E. J Biol Chem 281: Karin M, Ben-Neriah Y. (2000). Phosphorylation meets 20788–20800. ubiquitination: The control of NF-kappa B activity. Annu Musti AM, Treier M, Bohmann D. (1997). Reduced ubiquitin- Rev Immunol 18: 621–663. dependent degradation of c-Jun after phosphorylation by Kim TK, Maniatis T. (1996). Regulation of interferon-gamma- MAP kinases. Science 275: 400–402. activated STAT1 by the ubiquitin-proteasome pathway. Nimmanapalli R, Bali P, O’Bryan E, Fuino L, Guo F, Wu J Science 273: 1717–1719. et al. (2003). Arsenic trioxide inhibits translation of mRNA Krappmann D, Scheidereit C. (2005). A pervasive role of of bcr-abl, resulting in attenuation of Bcr-Abl levels ubiquitin conjugation in activation and termination of I and apoptosis of human leukemia cells. Cancer Res 63: kappa B kinase pathways. EMBO R 6: 321–326. 7950–7958. Kumar KGS, Krolewski JJ, Fuchs SY. (2004). Phosphoryla- Nishizawa M, Furuno N, Okazaki K, Tanaka H, Ogawa Y, tion and specific ubiquitin acceptor sites are required for Sagata N. (1993). Degradation of Mos by the N-terminal ubiquitination and degradation of the IFNAR1 subunit of proline (Pro2)-dependent ubiquitin pathway on fertilization type I interferon receptor. J Biol Chem 279: 46614–46620. of Xenopus eggs: possible significance of natural selection Kumar V, Sabatini D, Pandey P, Gingras AC, Majumder PK, for Pro2 in Mos. EMBO J 12: 4021–4027. Kumar M et al. (2000). Regulation of the rapamycin and Okazaki K, Sagata N. (1995). The Mos/MAP kinase pathway FKBP-target 1/mammalian target of rapamycin and cap- stabilizes c-Fos by phosphorylation and augments its dependent initiation of translation by the c-Abl protein- transforming activity in NIH 3T3 cells. EMBO J 14: tyrosine kinase. J Biol Chem 275: 10779–10787. 5048–5059. Le Bouffant R, Cormier P, Mulner-Lorillon O, Belle R. (2006). Orford K, Crockett C, Jensen JP, Weissman AM, Byers SW. Hypoxia and DNA-damaging agent bleomycin both increase (1997). Serine phosphorylation-regulated ubiquitination the cellular level of the protein 4E-BP. J Cell Biochem 99: and degradation of beta-catenin. J Biol Chem 272: 126–132. 24735–24738. Li BG, Fang CH, Hasselgren P. (2000). Degradation of Othumpangat S, Kashon M, Joseph P. (2005a). Eukaryotic methoxysuccinyl-phe-leu-phe-7-amido-4-trifluoromethyl translation initiation factor 4E is a cellular target for toxicity coumarin (FLF) in cultured myotubes and HepG2 cells is and death due to exposure to cadmium chloride. J Biol Chem proteasome- and calpain/calcium-dependent. Int J Biochem 280: 25162–25169. Cell Biol 32: 677–686. Othumpangat S, Kashon M, Joseph P. (2005b). Sodium Li S, Sonenberg N, Gingras AC, Peterson M, Avdulov S, arsenite-induced inhibition of eukaryotic translation initia- Polunovsky VA et al. (2002). Translational control of cell tion factor 4E (eIF4E) results in cytotoxicity and cell death. fate: availability of phosphorylation sites on translational Mol Cell Biochem 279: 123–131. repressor 4E-BP1 governs its proapoptotic potency. Mol Cell Paglin S, Lee NY, Nakar C, Fitzgerald M, Plotkin J, Deuel B Biol 22: 2853–2861. et al. (2005). Rapamycin-sensitive pathway regulates mito- Lin DI, Barbash O, Kumar KG, Weber JD, Harper JW, chondrial membrane potential, autophagy, and survival in Klein-Szanto AJ et al. (2006a). Phosphorylation-dependent irradiated MCF-7 cells. Cancer Res 65: 11061–11070. ubiquitination of cyclin D1 by the SCF(FBX4-alphaB Peterson RT, Desai BN, Hardwick JS, Schreiber SL. (1999). crystallin) complex. Mol Cell 24: 355–366. Protein phosphatase 2A interacts with the 70-kDa S6 Lin KI, Baraban JM, Ratan RR. (1998). Inhibition versus kinase and is activated by inhibition of FKBP12- induction of apoptosis by proteasome inhibitors depends on rapamycin-associated protein. Proc Natl Acad Sci USA 96: concentration. Cell Death Differ 5: 577–583. 4438–4442. Lin SS, Bassik MC, Suh H, Nishino M, Arroyo JD, Hahn WC Proud CG. (2002). Regulation of mammalian translation et al. (2006b). PP2A regulates BCL-2 phosphorylation and factors by nutrients. Eur J Biochem 269: 5338–5349. proteasome-mediated degradation at the endoplasmic Proud CG. (2005). The eukaryotic initiation factor 4E-binding reticulum. J Biol Chem 281: 23003–23012. proteins and apoptosis. Cell Death Differ 12: 541–546. Lin T-A, Kong X, Haystead TAJ, Pause A, Belsham G, Ren Q, Kari C, Quadros MR, Burd R, McCue P, Dicker AP Sonenberg N et al. (1994). PHAS-I as a link between et al. (2006). Malignant transformation of immortalized mitogen-activated protein kinase and translation initiation. HaCaT keratinocytes through deregulated nuclear factor Science 266: 653–656. kappa B signaling. Cancer Res 66: 5209–5215.

Oncogene Phosphorylation, ubiquitination and stability of 4E-BP1 A Elia et al 822 Rosenwald IB, Kaspar R, Rousseau D, Gehrke L, Leboulch P, Tee AR, Proud CG. (2000). DNA-damaging agents cause Chen JJ et al. (1995). Eukaryotic translation initiation factor inactivation of translational regulators linked to mTOR 4E regulates expression of cyclin D1 at transcriptional and signalling. Oncogene 19: 3021–3031. post-transcriptional levels. J Biol Chem 270: 21176–21180. Tilleray V, Constantinou C, Clemens MJ. (2006). Regulation Rousseau D, Kaspar R, Rosenwald I, Gehrke L, Sonenberg N. of protein synthesis by inducible wild-type p53 in human (1996). Translation initiation of ornithine decarboxylase and lung carcinoma cells. FEBS Lett 580: 1766–1770. nucleocytoplasmic transport of cyclin D1 mRNA are Topisirovic I, Guzman ML, McConnell MJ, Licht JD, increased in cells overexpressing eukaryotic initiation factor Culjkovic B, Neering SJ et al. (2003). Aberrant eukaryotic 4E. Proc Natl Acad Sci USA 93: 1065–1070. translation initiation factor 4E-dependent mRNA transport Salau¨ n P, Pyronnet S, Morales J, Mulner-Lorillon O, Belle´ R, impedes hematopoietic differentiation and contributes to Sonenberg N et al. (2003). eIF4E/4E-BP dissociation and leukemogenesis. Mol Cell Biol 23: 8992–9002. 4E-BP degradation in the ﬁrst mitotic division of the sea Torres C, Francis MK, Lorenzini A, Tresini M, Cristofalo VJ. urchin embryo. Dev Biol 255: 428–439. (2003). Metabolic stabilization of MAP kinase phosphatase- Scheidereit C. (2006). I kappa B kinase complexes: gateways to 2 in senescence of human ﬁbroblasts. Exp Cell Res 290: NF-kappa B activation and transcription. Oncogene 25: 195–206. 6685–6705. von der Haar T, Gross JD, Wagner G, McCarthy JEG. (2004). Schneider R, Braunstein S, Xi Q, Formenti S. (2005). Ionizing The mRNA cap-binding protein eIF4E in post-transcrip- radiation controls protein synthesis through a novel tional gene expression. Nat Struct Biol 11: 503–511. Akt-independent pathway involving regulation of mTOR Walsh D, Mohr I. (2004). Phosphorylation of eIF4E by Mnk-1 and 4E-BP1 stability. Int J Radiat Oncol Biol Phys 63: S146. enhances HSV-1 translation and replication in quiescent Song Q, Lavin MF. (1993). Calyculin A, a potent inhibitor of cells. Genes Dev 18: 660–672. phosphatases-1 and -2A, prevents apoptosis. Biochem Walsh D, Perez C, Notary J, Mohr I. (2005). Regulation of the Biophys Res Commun 190: 47–55. translation initiation factor eIF4F by multiple mechanisms Strudwick S, Borden KLB. (2002). The emerging roles of in human cytomegalovirus-infected cells. J Virol 79: translation factor eIF4E in the nucleus. Differentiation 70: 8057–8064. 10–22. Wan X, Mendoza A, Khanna C, Helman LJ. (2005). Tee AR, Blenis J. (2005). mTOR, translational control and Rapamycin inhibits ezrin-mediated metastatic behavior in human disease. Semin Cell Dev Biol 16: 29–37. a murine model of osteosarcoma. Cancer Res 65: 2406–2411.

Oncogene