cdc2–cyclin B regulates eEF2 kinase activity in a cell cycle- and amino acid-dependent manner
2008; Springer Nature; Volume: 27; Issue: 7 Linguagem: Inglês
10.1038/emboj.2008.39
ISSN1460-2075
AutoresEwan M. Smith, Christopher G. Proud,
Tópico(s)Polyamine Metabolism and Applications
ResumoArticle13 March 2008free access cdc2–cyclin B regulates eEF2 kinase activity in a cell cycle- and amino acid-dependent manner Ewan M Smith Ewan M Smith Division of Molecular Physiology, College of Life Sciences, University of Dundee, Dundee, UK Department of Biochemistry & Molecular Biology, Life Sciences Centre, University of British Columbia, Vancouver, British Columbia, Canada Search for more papers by this author Christopher G Proud Corresponding Author Christopher G Proud Division of Molecular Physiology, College of Life Sciences, University of Dundee, Dundee, UK Department of Biochemistry & Molecular Biology, Life Sciences Centre, University of British Columbia, Vancouver, British Columbia, Canada Search for more papers by this author Ewan M Smith Ewan M Smith Division of Molecular Physiology, College of Life Sciences, University of Dundee, Dundee, UK Department of Biochemistry & Molecular Biology, Life Sciences Centre, University of British Columbia, Vancouver, British Columbia, Canada Search for more papers by this author Christopher G Proud Corresponding Author Christopher G Proud Division of Molecular Physiology, College of Life Sciences, University of Dundee, Dundee, UK Department of Biochemistry & Molecular Biology, Life Sciences Centre, University of British Columbia, Vancouver, British Columbia, Canada Search for more papers by this author Author Information Ewan M Smith1,2 and Christopher G Proud 1,2 1Division of Molecular Physiology, College of Life Sciences, University of Dundee, Dundee, UK 2Department of Biochemistry & Molecular Biology, Life Sciences Centre, University of British Columbia, Vancouver, British Columbia, Canada *Corresponding author. Department of Biochemistry & Molecular Biology, University of British Columbia, Life Sciences Centre, 2350 Health Sciences Mall, Vancouver, British Columbia, Canada V6T 1Z3. Tel.: +1 604 827 3923; Fax: +1 604 822 5227; E-mail: [email protected] The EMBO Journal (2008)27:1005-1016https://doi.org/10.1038/emboj.2008.39 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The calcium/calmodulin-dependent kinase that phosphorylates and inactivates eukaryotic elongation factor 2 (eEF2 kinase; eEF2K) is subject to multisite phosphorylation, which regulates its activity. Phosphorylation at Ser359 inhibits eEF2K activity even at high calcium concentrations. To identify the kinase that phosphorylates Ser359 in eEF2K, we developed an extensive purification protocol. Tryptic mass fingerprint analysis identified it as cdc2 (cyclin-dependent kinase 1). cdc2 co-purifies with Ser359 kinase activity and cdc2–cyclin B complexes phosphorylate eEF2K at Ser359. We demonstrate that cdc2 contributes to controlling eEF2 phosphorylation in cells. cdc2 is activated early in mitosis. Kinase activity against Ser359 in eEF2K also peaks at this stage of the cell cycle and eEF2 phosphorylation is low in mitotic cells. Inactivation of eEF2K by cdc2 may serve to keep eEF2 active during mitosis (where calcium levels rise) and thereby permit protein synthesis to proceed in mitotic cells. Amino-acid starvation decreases cdc2's activity against eEF2K, whereas loss of TSC2 (a negative regulator of mammalian target of rapamycin complex 1(mTORC1)) increases it. These data closely match the control of Ser359 phosphorylation and indicate that cdc2 may be regulated by mTORC1. Introduction Eukaryotic elongation factor 2 (eEF2) mediates the translocation step of peptide-chain elongation (Merrick and Nyborg, 2000). Phosphorylation of Thr56 in eEF2 prevents eEF2 from interacting with the ribosome, thereby inactivating it (Carlberg et al, 1990). eEF2 is phosphorylated by eEF2 kinase (eEF2K) (Nairn et al, 1985; Ryazanov et al, 1988). The phosphorylation of eEF2 is modulated under a range of conditions, suggesting that eEF2K is a key regulator of translation elongation (Herbert and Proud, 2006) and, thus, protein synthesis. eEF2K is an unusual calcium/calmodulin (Ca/CaM)-dependent protein kinase. Consequently, conditions associated with elevated calcium levels, such as muscle contraction, are associated with a rapid increase in eEF2 phosphorylation (Rose et al, 2005). By slowing translation elongation, this may conserve energy for use by the contractile machinery. However, there are other situations, such as mitosis, where calcium levels rise but translation is inhibited at initiation rather than elongation, and where activation of eEF2K therefore appears inappropriate (Fan and Penman, 1970; Pyronnet and Sonenberg, 2001; Whitaker, 2006; Wilker et al, 2007). Indeed, specific mRNAs continue to be translated during mitosis and it is thus important to keep the elongation machinery active. Stimuli that activate protein synthesis elicit the inactivation of eEF2K and dephosphorylation of eEF2 (reviewed in Herbert and Proud (2006); Wang and Proud (2006)). In the case of insulin, both effects are blocked by rapamycin, which interferes with some functions of the mammalian target of rapamycin complex 1 (mTORC1) (Wullschleger et al, 2006), indicating that insulin's effects on eEF2K are mediated through mTORC1. However, it should be noted that not all the effects of mTORC1 are blocked by rapamycin. For example, the phosphorylation of certain regulatory sites (Thr37/46) in eukaryotic initiation factor 4E-binding protein 1 (4E-BP1) is quite insensitive to this drug (Gingras et al, 2001; Wang et al, 2005), even though multiple lines of evidence indicate that they are controlled by mTORC1. Indeed, mTORC1 can directly phosphorylate these sites in vitro, albeit in a quite rapamycin-insensitive manner (McMahon et al, 2002). Their phosphorylation is profoundly decreased by depriving cells of amino acids, especially leucine, which strongly impairs mTORC1 signalling. Other targets for mTORC1, such as the S6 kinases, are strongly affected both by amino-acid starvation and by rapamycin (Avruch et al, 2001). eEF2K is subject to phosphorylation in vivo at several sites (Browne and Proud, 2002; Wang and Proud, 2006). Phosphorylation of eEF2K at certain sites decreases its activity, whereas phosphorylation at others increases it (Herbert and Proud, 2006). The phosphorylation of eEF2K at Ser359 is of particular interest. First, the phosphorylation of this site strongly decreases the activity of eEF2K even at high calcium concentrations (Knebel et al, 2001), that is, desensitizes eEF2K to the activating effects of elevated Ca2+ levels. Second, Ser359 is partially phosphorylated under basal conditions. The phosphorylation of this site is decreased by starving cells of amino acids, suggesting that mTORC1 may have an important function in controlling this site. Third, although p38 MAP kinase δ (also termed SAPK4δ) can phosphorylate Ser359 in vitro (Knebel et al, 2001), this enzyme is not known to be active basally or to be regulated by amino acids. It therefore appeared likely that Ser359 was also phosphorylated by another protein kinase, which might be controlled by the amino-acid status of the cells and perhaps by mTORC1. Our aim was to identify this kinase, a potential new link to the control of translation elongation. Results Development of an assay for kinase activity against Ser359 in eEF2K To allow us to easily and specifically monitor kinase activity against Ser359 in eEF2K, we needed to develop an assay procedure to detect activity against this site. There were two reasons for this. First, eEF2K itself undergoes (auto)phosphorylation (Redpath and Proud, 1993) and the resulting incorporation of radiolabel into eEF2K presents major problems for detecting the activity of eEF2K. Secondly, eEF2K undergoes phosphorylation at multiple sites and we wished to identify the kinase that specifically phosphorylates Ser359. Our assay exploits a new phospho-specific antibody for Ser359. Its characterization is documented in Supplementary Figure S1. Neither the crude serum (Supplementary Figure S1A) nor the affinity-purified antibody (Supplementary Figure S1B) reacted with the non-phosphorylated peptide, although a strong signal was seen with its phospho-version, defining the antibody as truly phospho-specific. To assess whether Ser359 kinase activity could be detected in KB cell lysates, we incubated lysate from serum-starved cells with GST–eEF2K and ATP. Reaction products were analysed by SDS–PAGE/western blot using the Ser359-phospho-specific antibody. Kinase activity against Ser359 was detected in such lysates (Supplementary Figure S1C) but not in lysates of amino acid-starved cells, consistent with the loss of phosphorylation of Ser359 that occurs upon amino-acid deprivation (Browne and Proud, 2004). These data show that the assay detects Ser359 kinase activity whose behaviour matches that expected for a physiologically relevant eEF2K Ser359 kinase. It was first important to establish whether two potential candidates, mTORC1 and SAPK4 (Knebel et al, 2001), accounted for this activity. eEF2K is not directly phosphorylated by mTOR and does not bind raptor It was possible that mTORC1, signalling through which is positively regulated by amino acids, directly phosphorylated eEF2K. To test this, we performed immunoprecipitations using anti-raptor (for mTORC1), anti-rictor (mTORC2) or anti-mTOR (mTAb1; for both complexes). When any of the mTOR complexes was incubated with eEF2K, no phosphorylation of Ser359 in eEF2K was observed (Figure 1A). In contrast, robust phosphorylation of a known mTORC1 substrate, 4E-BP1, was seen where expected (Figure 1B). These data demonstrate that the regulation of Ser359 in eEF2K by mTORC1 is indirect, implying the existence of an additional Ser359 kinase. Figure 1.Neither mTORC1 nor SAPK4 is the amino acid-regulated eEF2K Ser359 kinase. (A, B) KB cells were harvested in CHAPS lysis buffer. mTOR was immunoprecipitated with indicated antibody bound to protein G-Sepharose. Immunoprecipitates were washed three times in lysis buffer and then used in kinase assays with (A) eEF2K (right-hand lane is a control using SAPK4 to phosphorylate eEF2K) or (B) 4E-BP1 (positive control) as substrate. Assay products were subjected to SDS–PAGE/western blot analysis with the indicated antibodies. (C) Raptor does not bind eEF2K: recombinant GST 4E-BP1, eEF2K and eIF2α were subjected to SDS–PAGE, and then transferred onto PVDF membrane for ‘far western analysis’. The blot was incubated with lysates from HEK293 cells expressing myc-raptor and probed with anti-myc. The membrane was then stripped and re-probed using anti-GST. (D) HEK293 cells were transiently transfected with DNA for myc-SAPK4. At 24 h later, cells were starved of serum for 16 h. Cells were then treated as indicated above (*D-PBS was supplemented with amino acids). Myc-SAPK4 was immunoprecipitated from HEK293 cell lysates and activity was measured versus GST–eEF2K. Kinase assay products were analysed by SDS–PAGE/western blot using the indicated antibodies. Download figure Download PowerPoint To examine why eEF2K is not phosphorylated by mTORC1, we examined whether it can bind to the mTORC1 component, raptor, as is the case for direct mTORC1 substrates such as 4E-BP1, which contain TOR-signalling motifs (Schalm and Blenis, 2002). Using our overlay assay (Beugnet et al, 2003), we did not observe any interaction of eEF2K with raptor, or eIF2α, a negative control (Figure 1C). Raptor did bind to 4E-BP1, a positive control. Thus, the inability of mTORC1 to phosphorylate eEF2K may reflect the fact that, unlike 4E-BP1, eEF2K cannot bind raptor. Ser359 kinase activity in amino acid-fed cells is not due to SAPK4 SAPK4 (Knebel et al, 2001) is also a candidate for the Ser359 kinase activity observed in amino acid-fed cells. However, no activity whatsoever was observed in SAPK4 immunoprecipitates from such cells, even after insulin treatment. Strong SAPK4 activity was, however, detected after treatment of cells with anisomycin, which activates SAPK4 (Figure 1D; last lane; see Knebel et al (2001)). Purification and identification of the Ser359 kinase Purified kinases can often phosphorylate on non-physiological substrates in vitro. Therefore, to try to ensure accurate identification of the kinase responsible for phosphorylating Ser359, we elected to attempt to purify this enzyme(s) from lysates of KB cells, which are a relatively rich source of this activity. We developed a series of purification steps using anion and cation exchangers; HiTrap Blue, a nucleotide-based affinity column; and gel filtration. At each stage, all the Ser359 kinase activity was retained on the column (data for Resource Q, Figure 2A and B; also data not shown) and eluted as a single peak, indicating that there is one major species of Ser359 kinase in amino acid-fed KB cells. Extensive work showed that a combination of Resource Q, Mono S, HiTrap Blue and Superose 6 columns gave the best enrichment and recovery of activity (Figure 2C). The data in Figure 2A and B (derived from Resource Q ion-exchange chromatography of cell lysates) provide further confirmation that basal kinase activity is not due to SAPK4: Ser359 kinase activity eluted mainly in fractions 6–10, whereas SAPK4 appeared earlier and no SAPK4 protein was detected in the most active fractions (Figure 2B, lower part). Figure 2.Purification of eEF2K Ser359 kinase from KB cell lysates. (A, B) Resource Q FPLC was performed and 1-ml fractions were collected. (A) The elution profile, that is, the absorbance at 280 nm and the ionic strength of the elution buffer, is shown. (B) Western blot of kinase assay products (probed with anti-(P)Ser359 eEF2K) for the indicated fractions, the load and flowthrough. Samples were also analysed by western blot for SAPK4, as indicated. (C) Scheme outlining the large-scale purification procedure for the Ser359 kinase. (D) The activity of the eEF2K Ser359 kinase decreases upon incubation with protein phosphatase PP2A. Purified kinase activity from the Superose 6 gel filtration fraction 13 was incubated at room temperature for 15 min with or without 50 mU/ml of PP2A. Phosphatase activity was stopped by adding 1 μM microcystin. Kinase assays were then performed against GST–eEF2K as described in Materials and methods section. ‘Mock’—kinase was pre-incubated at 4°C prior to assay. *Phosphatase was inactivated by adding 1 μM microcystin prior to incubation with Ser359 kinase. Download figure Download PowerPoint The activity of the highly purified eEF2K Ser359 kinase decreased after incubation with the Ser/Thr phosphatase PP2A (Figure 2D), indicating that the Ser359 kinase is activated by phosphorylation and implying that it is a phosphoprotein. To identify the kinase, we therefore focused our attention on any phosphoproteins that were present in the purified material: the proteins in the most active fractions from the gel filtration column were precipitated with TCA and analysed by SDS–PAGE. Application of the phosphoprotein stain ProQ Diamond® (Supplementary Figure S2) revealed several bands. Each band was excised, digested with trypsin and analysed by mass spectrometry. The data (Table I) revealed multiple peptides from the cell division control protein cdc2 (also termed cyclin-dependent kinase 1 (CDK1)), but no peptides from other protein kinases. Altogether, 20 peptides matched cdc2 (covering >50% of its sequence), including one phosphopeptide. This contained phospho-Thr161, phosphorylation of which is necessary for cdc2 to be active (Solomon et al, 1993). This may account for the phosphatase-sensitivity of the purified Ser359 kinase (Figure 2D). Table 1. Mascot search results Band ID Accession Protein Peptides matched 1 P06493 Niban-like protein (Meg3) 19 O43583 Glutaminyl-tRNA synthetase 15 2 P02768 Serum albumin precursor 9 3 P49748 Very-long chain-specific acyl-coA dehydrogenase, mitochondrial precursor 7 P50579 Methionine aminopeptidase 2 6 4 Q9UQ80 Proliferation-associated protein 2G4 24 P38919 Probable ATP-dependent helicase DDX48 7 5 P064493 Cell division control protein 2 20 O43583 Density-regulated protein 8 6 Q9UQ80 Proliferation-associated protein 2G4 25 P38919 Probable ATP-dependent helicase DDX48 9 The peptide masses for tryptic peptides from excised gel bands (see Supplementary Figure S2) were subjected to analysis by MS. Data presented show the first two significant hits from the Mascot (http://matrixscience.com) database searches performed on resulting peptide masses of each band (1–6). cdc2–cyclin B phosphorylates eEF2K at Ser359 in vitro and the CDK inhibitor roscovitine decreases Ser359 phosphorylation in vivo Western blot analysis revealed that active fractions from the Resource Q column contained cdc2 and cyclin B (Figure 3A). Ser359 kinase activity also co-purified with cdc2 during the second purification step (Mono S; Figure 3B) and at all subsequent stages (data not shown), consistent with the identification of the eEF2K Ser359 kinase as cdc2. The most highly purified active fractions also contained cyclin B (Figure 3C). Figure 3.Ser359 kinase activity co-purifies with cdc2–cyclin B. (A) Resource Q; (B) Mono S and (C) gel filtration FPLC were performed on lysates from insulin-treated KB cells (see Supplementary data). Fractions were assayed for kinase activity and products were subjected to SDS–PAGE/western blotting with indicated antisera (F/T, flowthrough). Download figure Download PowerPoint The cdc2 inhibitor roscovitine completely inhibited Ser359 kinase activity both in the peak fraction from the Resource Q column and in KB cell lysates (Figure 4A and B), indicating that no other, roscovitine-insensitive, kinase(s) contribute to this activity. The Resource Q-purified Ser359 kinase also phosphorylated histone H1 (a well-established substrate for cdc2–cyclin B; see, e.g., Clarke et al (1992)). Figure 4.Ser359 kinase activity is inhibited by roscovitine. (A) Fraction 7 of Resource Q FPLC (Figure 3A) was assayed for kinase activity against eEF2K and histone H1 in the presence of 10 μM roscovitine or DMSO. Assay products were analysed by western blotting or autoradiography as indicated. (B) KB cell lysates or (C) cdc2 immunoprecipitates were assayed for activity towards Ser359 in eEF2K in the presence or absence of roscovitine (10 μM). Assay products were subjected to SDS–PAGE/western blot using the indicated antibodies. In (C), immunoprecipitates were also assayed against histone H1: in this case the figure is an autoradiograph of the stained gel. (D) Recombinant cdc2–cyclin B complexes were pre-incubated at room temperature for 20 min with 10 μM roscovitine (or DMSO as control). cdc2–cyclin B complexes were then incubated with 2 μg of GST–eEF2 kinase and ATP. Assay products were analysed by western blotting. (E) HEK293 cells were transfected with myc-eEF2K. 40 h later, cells were treated with DMSO or 50 μM roscovitine for 1 h prior to lysis. SDS–PAGE/western blotting was performed on myc immunoprecipitates, cell lysates or cell pellets as indicated. Download figure Download PowerPoint cdc2 immunoprecipitates from KB cell lysates phosphorylated eEF2K at Ser359 (Figure 4C) and recombinant cdc2–cyclin B also phosphorylated eEF2K at this site (Figure 4D). Activity was again completely blocked by roscovitine. These data show that cdc2 can indeed phosphorylate eEF2K at Ser359. The sequence around Ser359 in eEF2K is CGSPRVRTL, similar to the optimal consensus for phosphorylation by cdc2–cyclin B, S/T-P-X-K/R (X=any amino acid) (Ubersax et al, 2003), although CDKs can phosphorylate sites with the minimal consensus S/T-P. To study further the phosphorylation and regulation of eEF2K by cdc2, we first tested the effect of cdc2-mediated phosphorylation on its activity. Incubation of eEF2K with cdc2–cyclin B and ATP led to a marked decrease in its activity against eEF2 (Supplementary Figure S3A), consistent with earlier data (Knebel et al, 2001) that phosphorylation at Ser359 inhibits eEF2K activity. To test whether Ser359 is the major cdc2 site in eEF2K, we mutated Ser359 to Ala (using a newly created kinase-dead version of eEF2K (Lys170Met; to be reported elsewhere) to avoid complications arising from the propensity of eEF2K to autophosphorylate (Redpath and Proud, 1993). This mutation markedly decreased, but did not entirely abolish, cdc2-catalysed phosphorylation of eEF2K (Supplementary Figure S3B), showing that, although Ser359 is the major site for cdc2, it is not the only one. The S359A mutation abolished the ability of cdc2 to inactivate eEF2K (Supplementary Figure S3A), showing that the minor cdc2 sites do not affect eEF2K's activity. Scansite (http://scansite.mit.edu/) does predict a second potential cdc2 site in eEF2K at Ser329, albeit with lower confidence than Ser359. We therefore also mutated this residue to alanine. The S359A/S329A mutant showed no further decrease in phosphorylation by cdc2 (data not shown), indicating that additional sites are phosphorylated, at least in vitro. It was crucial to establish whether cdc2 actually regulates the phosphorylation of eEF2K and its substrate, eEF2, in cells. Because the Ser359 phospho-specific antibody was insufficiently sensitive to detect endogenous eEF2K (data not shown), we overexpressed (myc)-eEF2K, using HEK293 cells (because they can be transfected more efficiently than KB cells). After immunoprecipitation, a signal for phosphorylated Ser359 in eEF2K was clearly seen in control cells (Figure 4E). This signal was completely lost if cells were pretreated with roscovitine (for only 1 h, to avoid affecting cell cycle progression), consistent with the conclusion that cdc2 is the only kinase that phosphorylates this site within cells. Furthermore, roscovitine treatment increased the phosphorylation of eEF2 (Figure 4E). These data provide strong evidence that Ser359, an inhibitory site in eEF2K, is regulated by cdc2 in vivo and that this is important for the control of eEF2K activity and eEF2 phosphorylation. Analysis of the phosphorylation of endogenous histone H1 (a cdc2 substrate) confirmed the efficacy of roscovitine (Figure 4E). Regulation of eEF2K and eEF2 phosphorylation during the cell cycle There is substantial evidence that calcium transients have an important function during mitosis in many types of cells, including mammalian cells (FitzHarris et al, 2005; Whitaker, 2006). Indeed, calcium (through its binding to CaM) apparently has an important function in mitotic progression (Rasmussen and Means, 1989; Baitinger et al, 1990; Torok et al, 1998). Phosphorylation of eEF2K at Ser359 strongly inhibits its activity even at high (supraphysiological) concentrations of Ca2+ (10 μM) (Knebel et al, 2001). The finding that cdc2 phosphorylates eEF2K at Ser359 could therefore be important for the control of eEF2K, eEF2 and translation elongation during M-phase. To study this further, HeLa cells were blocked in early S-phase using the DNA polymerase inhibitor aphidicolin (Oguro et al, 1979). At various times after release from this block, progress through the cell cycle was monitored by flow cytometry (Figure 5A). Cells were also lysed and samples were analysed for kinase activity against Ser359 in eEF2K or by SDS–PAGE/western blot for other proteins (Figure 5B). Figure 5.Regulation of eEF2K Ser359 kinase activity during the cell cycle. HeLa cells were synchronized in late G1–S-phase by a 16 h aphidicolin block (or DMSO as control) and then released into fresh medium for various times. (A) Cell cycle progression was monitored by flow cytometry of propidium iodide-stained cells at the indicated intervals. Series indicate G1 (2N), S and G2/M (4N) DNA content. Results show mean±s.d. for three independent experiments. (B) Cell extracts were prepared from synchronized HeLa cells at the indicated times after release. In vitro kinase assays versus recombinant GST–eEF2K were performed on cdc2 immunoprecipitates from cell lysates. Assay products were subjected to SDS–PAGE and blotted with indicated antisera. Cell lysates were also subjected to SDS–PAGE and probed for eEF2 (P)Thr56, total eEF2. (C) Cell lysates from each time point after aphidicolin release were incubated with GST–eEF2K and ATP, ±10 μM roscovitine. SDS–PAGE/western blot analysis of assay products was performed using the (P)Ser359 eEF2K antibody. (D) As (B) but cell extracts were subjected to SDS–PAGE/western blot for S6 (P)235/236 and total S6. Download figure Download PowerPoint Flow cytometry revealed that the proportion of G2/M cells was maximal 8 h after release. The proportion of G1 cells increased at 10–12 h (indicating exit from mitosis) and cells began to re-enter S-phase at 18 h. Ser359 kinase activity in cdc2 immunoprecipitates was undetectable immediately after release but then rose, peaking at 8–10 h, when cyclin B levels were highest (Figure 5B). Total cdc2 levels were constant throughout (Figure 5B). The Ser359 kinase activity appears to ‘lag behind’ the proportion of ‘G2+M’ cells, presumably because cdc2 is only activated at late times when more cells are in M-phase and not during the preceding G2-phase. At all time points, Ser359 kinase activity (measured in cell lysates) was blocked by roscovitine (Figure 5C). The phosphorylation of endogenous eEF2 was high immediately after removing the block but low through G2 and mitosis, only rising again well after cells had reached G1 and began to re-enter S-phase (Figure 5B). Phosphorylation of S6, another target for mTORC1 signalling, remained high in G2/M and fell as cells re-entered G1 and progressed into S-phase (Figure 5D). To further substantiate the control of eEF2 and eEF2K Ser359 kinase activity during the cell cycle, we considered it important to employ a second, different, way to synchronize cells, blocking with thymidine. The data were very similar to those from the aphidicolin block (Supplementary Figure S4A; see also Figure 5B): Ser359 kinase activity was again maximal 8–9 h after release, when the proportion of cells in mitosis appears highest (Supplementary Figure S4B), as is kinase activity against histone H1. The data also confirm that eEF2 phosphorylation is very low during mitosis (Supplementary Figure S4A). Importantly, these data demonstrate that (i) eEF2K Ser359 kinase activity is greatest when one would expect cdc2 activity to be highest and (ii) in accordance with the fact that phosphorylation at Ser359 inhibits eEF2K activity even at high Ca2+ concentrations, eEF2 phosphorylation is low in mitotic cells. Phosphorylation of any of three sites in eEF2K can inactivate it, that is, Ser78 (Browne and Proud, 2004), Ser359 (Knebel et al, 2001) or Ser366 ((Wang et al, 2001) albeit only at low Ca2+ concentrations). We therefore measured kinase activity against these specific sites in eEF2K in lysates of cells sampled at different points in the cell cycle. Activity against Ser78 and Ser366 did not vary, that against Ser78 being low at all times tested (Supplementary Figure S4C). Ser366 is a substrate for S6 kinases and S6 phosphorylation was also similar at all times after release from thymidine block. In contrast, activity against Ser359 did change, being low early after release and highest at the times when the proportion of G2+M cells was highest (5 and 9 h). eEF2 phosphorylation was lowest at these times, although it did also decrease at 3 h (both after release from thymidine (Supplementary Figure S4A and C) or aphidicolin (data not shown). At this time point, only a small activation of Ser359 kinase was evident. eEF2K is subject to an array of regulatory inputs and additional factors may explain this decrease. Regulation of the Ser359 kinase activity of cdc2 by amino acids Previous work has shown that amino-acid starvation and rapamycin each decrease the phosphorylation of Ser359 in eEF2K (Knebel et al, 2001; Browne and Proud, 2004), implying that they may affect the activity of the relevant kinase, which we have identified as cdc2. As the above data indicate that the Ser359 kinase activity that can be detected in cell lysates is due to cdc2, we examined the effects of amino-acid starvation on this activity. When KB cells were starved of amino acids, Ser359 kinase activity declined slightly by 1.5 h, markedly by 3 h and activity was almost completely lost by 5 h (Figure 6A). This decrease could be due to inactivation of the kinase or to its degradation. To study this, we used the proteasome inhibitor MG132. As expected, MG132 increased levels of c-myc, which is degraded by the proteasome (Figure 6A). However, MG132 actually enhanced the decrease in Ser359 kinase activity seen in amino acid-starved cells. The decline in Ser359 kinase activity is therefore not due to its proteasomal degradation, but to other mechanisms. Proteasome activity may help maintain intracellular amino-acid levels and thus mTORC1 signalling: indeed, MG132 exacerbates the effect of amino-acid starvation (Vabulas and Hartl, 2005). Figure 6.Effect of amino-acid deprivation or TSC2 status on cdc2 activity. (A) KB cells were starved of serum for 12 h. Then the medium was changed to EBSS lacking amino acids (‘−AA’) or containing amino acids (‘+AA’) for the times indicated. The proteasome inhibitor MG132 (20 μM) or DMSO (−) was present for the duration of the starvation. Cells were lysed and Ser359 kinase activity was assayed using recombinant GST–eEF2K as substrate. Assay products and cell lysates were analysed by SDS–PAGE/western blotting using the indicated antisera. (B, C) KB cells were starved of serum for 14 h and subsequently of amino acids (where indicated) for 5 h (*amino acids were present in the EBSS
Referência(s)