- Email: [email protected]
Casein kinase-2 phosphorylates serine-2 in the β-subunit of initiation factor-2
Biochimica et Biophysica Acta, 1010 (1989) 377-380 Elsevier
nnA 1O230
377 BBA Report
Casein ldnase-2 phosphorylates serine-2 in the fi-subunit of initiation factor-2 S t e p h a n i e J. Clark, A n t h o n y J. A s h f o r d , N i g e l T. Price a n d C h r i s t o p h e r G. P r o u d Department of Biochemistry, School of Medical Sciences, University of Bristol, Bristol (U.K.)
(Received 10 August1988)
Key words: Caseinkinase-2; Initiationfactor; Proteinphosphorylation;Proteinsynthesis We have previously presented evidence which suggests that casein Hnase-2 phosphofflates a serine residue near the N-terminus of the ~-subunit of the initiation |actor elF-2 (Clark, S.J. et al. Biochim. Biophys. Acta 968, 2 | | - 2 1 9 ) . We now report further data which confirm that it is serine-2 which is phosphory|ated by casein kinase-2. This data includes (1) the electrophoretic mobi|ities of the phosphopepfides produced by different cleavage techniques, (2) the amino acid composition of the principal phosphopepfide generated by treatment with cyanogen bromide and (3) the resistance of this phosphopepfide to Edman degradation.
Initiation factor-2 (elF-2) mediates the binding of the initiator methionyl-tRNA (Met-tRNAi) to the 40 S ribosome subunit during peptide chain initiation ia eukaryotic cells, eIF-2 is a heterotrimeric protein composed of subunits termed a (eIF-2a, molecular mass 37 kDa),/~ (eIF-2/~, which migrates anomalously on SDSPAGE: in our system it exhibits an apparent molecular mass of about 54 kDa) and y (eIF-2~,, apparent molecular mass 50 kDa). eIF-2a can be phosphorylated at serine-51 by at least two distinct protein kinases, phosphorylation at this site leads to inhibition of peptidechain initiation by impairing the recycling of elF-2 between successive rounds of initiation (reviewed in Refs. 1-3). eIF-2/t can also be phosphorylated, on serines, by several protein kinases including casein kinase-2 [4-6] and protein kinase C [7]. Phosphorylation by these kinases occurs at different sites which appear to be adjacent in the primary sequence of eIF-2fl [8]. Changes in the level of phosphorylation of eIF-2fl accompany the changes in rates of peptide-chain initiation which result from heat-shock or serum depletion in HeLa cells [9,~0]. Here we present data showing that serine-2 of eIF-2fl is phosphorylated by casein kinase-2. Chemicals and biochemicals were all obtained from the sources described in ref. 8. The proteinase from
Abbreviations" elF-2, eukaryoticinitiationfactor-2; elF-2a, -2/~ and -2y, a-, /I- and ~,-subunits, respectively, of eukaryotic initiation factor-2. Correspondence: C.G. Proud, Departmentof Biochemistry,Schoolof Medical Sciences, Universityof Bristol, Bristol BS8 1TD, U.K.
Pseudomonas fragi which cleaves peptide bonds on the C-terminal side of aspartic acid residues was a kind gift from Dr. G.R. Drapeau, University of Montreal. Casein kinase-2 was prepared from rat liver and assayed as described previously [8]. elF-2 was isolated from rabbit reticulocytes as described in Ref. 11. elF-2 was phosphorylated by casein kinase-2 (0.5 U per #g elF-2) at 30 °C for 40 rain in reaction mixtures containing 3-~Nmorpholino)propane sulphonic acid (50 mM, pH 7.0), MgCI 2 (1 mM), KCI (150 raM), fl-mercaptoethanol (7.5 raM) and [y-32p]ATP (0.1 mM spec. radioact. 0.5 Ci. retool-t). Reactions were terminated by boiling in SDS sample buffer prior to loading onto SDS-polyacrylamide gels. Gels contained 12.5~ (w/v) acrylamide and 0.13~ (w/v) methylene bis-acrylamide as described previously [12] and, under these conditions, eIF-2fl is the slowest migrating subunit. After fixing, staining and destaining the gel, the band corresponding to eIF-2fl was excised and the fl-subunit was electroeluted in buffer containing 30 mM Tris-HC! (pH 7.8), 2 mM EDTA, 0.05~ (w/v) SDS and 15 mM fl-mercaptoethanol [13]. The eluate was dialysed for 6 h against 50 mM ammonium hydrogen carbonate, pH 7.8, and then concentrated in a Speed Va¢ vacuum concentrator (Savant, Hicksville, NY, USA). Protein was precipitated by adding acetone (4 volumes) at - 8 0 ° C for 1 h. The precipitated protein was then dissolved in either 50 mM ammonium hydrogen carbonate (pH 7.8) (for proteinase digestion) or 70~ (v/v) formic acid (for cyanogen bromide digestion). Aliquots (50 pl) of redissolved peptide in 50 mM ammonium hydrogen carbonate (pH 7.8) were treated
0167-4889/89/$03.50 © 1989ElsevierSciencePubfishersB.V.(BiomedicalDivision)
378 with Staphylococcus aureus B8 proteinase (10 ~g. m1-1) or P. fragi proteinase (10/tg. ml -]) for 2 h at 37°C, followed by addition of an equal amount of fresh proteinase and incubation was continued for a further 18 h. Other samples were digested with cyanogen bromide (100 mg. ml -~) in a final concentration of 70~o (v/v) formic acid [8]. After digestion, samples were dried in the vacuum concentrator and then dissolved either in 0.1~ (v/v) trifluoroacetic acid (for reverse-phase chromatography) or in 50~ (v/v) formic acid (for application onto thin-layer plates) [8]. Amino acid analysis was performed for us by Dr. David Campbell at the Department of Biochemistry, University of Dundee. After hydrolysis for 16 h at I I 0 ° C in 6 M HCI/2 mM phenol, samples were analysed using a Waters Picotag system as described in Ref. 14. The N-terminal sequence of eIF-2/~, obtained by eDNA cloning techniques [15] is shown below (Scheme I) (X indicates that the N-terminus is blocked [16], serines are shown bold and acidic residues are underlined). From the complete sequence of elF-2fl and the known substrate specificities of casein kinase-2 and protein kinase C, together with data showing that the sites in elF-2~ which are labelled by casein kinase-2 and protein kinase C are located in a single tryptic peptide, we have previously deduced that the phosphorylation sites are situated near the N-terminus of elF-2~ [8]. This tryptic l~eptide probably represents residues 1-14 of elF-2/~, and was previously reported to contain all the label intr,~aced by casein kinase-2. Within this peptide, ~erine-2 has adjacent to it, on the C-terminal side, two ~c~dic residues, Asp-4 and Glu-5: C-terminal acidic resm~.~ are the substrate specificity determinant for casein kinase-2 [17-20]. The N-terminal sequeace thus contains potential cleavage sites for S. aureus V8 proteinase (glutamic acid residues), P. fragi proteinase (aspartic acid residues) and also cyanogen bromide (methioni_ne residues). Treatment of elF-2/3 labelled by casein kinase-2 with cyanogen bromide, S. aureus V8 proteinase or P. fragi proteinase gave, in each case, a single anionic phosphopeptide (Fig. 1). Given that the N-terminus of elF-2/~ is apparently blocked [16], their electrophoretic properties are entirely consistent with their being derived from the N-terminal sequence of elF-2/~, i.e., X-Met-Ser(P)Gly-Asp-Glu-Met (where X represents the unknown blocking group). The peptides generated by each proce-
5
Q
Q
6 Fig. 1. Elec~rophoretic analysis at pH 3,6 of phosphopeptides derived from elF-2/3 labelled by casein kinase-2. The /3-subunit of elF-2 phosphorylated by casein kinase-2 was subjected to digestion by cyanogen bromide (lane 1), P. fragi proteinase (lane 2) or S. aureus V8 proteinase (lane 3) and the resulting peptides were analysed by electrophoresis at pH 3.6. The figure is an autoradiograph of the dried electrophoresis plate: the polarity ( + / - ) and the origin (arrow) are indicated. The square in lane 1 indicates the migration of the marker dinitrophenylaspartate.
dure would then be X-Met-Ser(P)-Gly-Asp-Glu-homoserine (cyanogen bromide cleavage product), X-MetSer(P-Gly-Asp-Glu (from V8 digestion) and X-MetSer(P)-Gly (generated by the P. fragi proteinase). The Met-Ser bond is clearly a possible additional site for cleavage by cyanogen bromide. However, the data shown in Fig. 1 (relative electrophoretic mobifities) and the amino acid composition data presented below show that cleavage did not occur at this bond. It is known that cyanogen bromide cleaves poorly at Met-Ser bonds [21]. On two occasions, a minor (less than 5% of total) radioactive species was observed migrating faster than the main cyanogen bromide digestion product (not shown): this may represent the X-Met-Ser(P) species. It did not represent another phosphorylation site located elsewhere in the sequence, since digestion of the labelled elF-2p with V8 or trypsin again only gave a single radiolabelled species (not shown, see also below). Based
10
15
X-Met-Sm'-Gly-Asl)-Glu-Met-Ile-Phe.As_.._~Pro.Th r.Met.Ser.Lys.Lys.Lys.Lys.Lys.Lys. Scheme I
379 Peok 20
1,¢
N
---
0 0 0 20 40 60 80 Fig. 2. Reverse-phase analysis of phosphopeptides generated by cyanogen bromide-digestion of elF-2fl phosphorylated by casein kinase-2. 32p radioactivity (o o); acetonitrile concentration (~). Horizontal bars labelled 1, 2 and 3 indicate the corresponding pools of radioactive material referred to in the text.
on the observation that the different cleavage procedures yidded peptides differing by only one or two amino acids in size but with markedly different electrophoretic mobilities (Fig. 1), the blocking group would appear to be small (of the order of the mass of an amino acid) and not highly charged. From the foregoing, it therefore seems that there is only a single major site for phosphorylation by casein kinase-2 in eIF-2fl, at least under our experimental conditions. Other workers have reported incorporation of up to two mol of phosphate per mall of enzymically dephosphorylated elF-2fl [5,22], but ~.~'.' have never observed stoichiometries greater than ~ mol of phosphate per mole of eIF-2fl. However, iF Jur more recent experiments, some of the label (usuahy less than 5%) remained at the origin during electrophoresis (Fig. 1). This differs from our earlier report, and may represent an additional, relatively minor, site for phosphorylation by casein kinase-2. This variability, in our hands, may
TABLE !
Amino acid composition of major phosphopeptidefrom elF-2fl labelled by casein kinase-2 All quantities have been corrected by subtraction of the values obtained when a similar volume of reverse-phase column buffer was analysed. Based on the associated radioactivity, approx. 130 pmol of phosphopeptide were analysed. Amino acid
Amount (pmol)
Aspartic acid Glutamic acid Serine Glycine Homoserine Methionine Alanine Valine All others
79 128 140 152 72 42 19 21 <8
be due to differing endogenous phosphate content in our purified eIF-2 preparations. The peptides resulting from cyanogen bromide cleavage of eIF-2fl labelled by casein kinase-2 were purified by reverse-phase chromatography on a C18 column [8]. The major radioactive peak e:ated at about 3% (v/v) acetonitrile (Fig. 2). Amino acid analysis of the peak fraction gave the composition shown in Table I. This amino acid composition of the phosphopeptide from eIF-2fl labelled by case:n kinase-2 is entirely consistent with phosphorylation at serine-2. The homoserine arises from the cyanogen bromide cleavage, whilst the methionine is probably present for reasons outlined above (i.e., it represents the N-terminal methionine). No precautions (other than carrying out the acid hydrolysis in vacuo) were taken to prevent oxidation of the methionine so the relatively poor recovery of this amino acid is not surprising. No sequence in the primary structure of elF-2fl, other than that at the N-terminus, could give rise to the composition obtained here [15]. On reverse-phase chromatography of the cyanogen bromide digest, two minor phosphopeptides were sometimes observed (Fig. 2). Together they accounted for 10-15% of the total radioactivity: one (peak 1) eluted slightly earlier from the C~8 column and the other eluted at about 5% (v/v) acetonitrile (peak 3). They may represent the products of cyanogen bromide cleavage at the first Met-Ser(P) bond giving Ser(P)-GlyAsp-Glu-Hser and at the Asp-Pro bond giving X-MetSer(P)-Gly-Asp-Glu-Met-Ile-Phe-Asp. Asp-Pro bonds are susceptible to hydrolysis in high concentrations of formic acid such as those (70% (v/v)) in which cyanogen bromide digestion is carried out [23]. The peptides in peaks 1 and 3 did not represent other phosphorylation sites in elF-2fl, since digestion of the same sample of phosphorylated elF-2/] by trypsin, V8-proteinase or P. fragi proteinase gave in each case only a single phosphopeptide species on electrophoresis or peptide mapping (not shown, but see Ref. 8). The major cyanogen bromide-generated phosphopeptide (peak 2) was refractory to the Edman degradation (result not shown). This observation indicates that it does indeed represent the N-terminal peptide of eIF-2fl, whose N-terminus is known to be blocked [16] and thus confirms that cyanogen bromide does not remove the N-terminal methionine from the major phosphopeptide species. As discussed above, casein kinase-2 phosphorylates serine or threonine residues which have several acidic amino acid residues immediately adjacent on the Cterminal side [17-20]. Casein kinase-2 generally 'prefers' substrates with larger clusters of acidic residues than the two adjacent to serine-2 in elF-2fl (Asp-4, Glu-5), but eIF-2fl does contain a glutamic acid residue at the particularly important +3 position [18]. Since serine-2 is the site phosphorylated by casein kinase-2
380
and since the casein kinase-2 and protein kinase C sites are located in the same tryptic peptide [8], the residue phosphorylated by protein kinase-C must be serine-13 as predicted from the data presented in Ref. 8. The activity of casein kinase-2 is known to undergo changes under a variety of physiological conditions and appears, for example, to be activated in response to insulin and epidermal growth factor [24-28]. It will, therefore, be of interest to discover whether the alterations in the level of phosphorylation of eIF-2~ which have been observed in response to various stimuli in HeLa cells [9,10] are occurring at the site phosphorylated by casein kinase-2. Having now identified the site labelled by casein kinase-2, it should be possible to investigate this. Although it is unclear how, if at all, phosphorylation of the/3-subunit affects eIF-2 activity, a recent report has suggested that the/~-subunit may be involved in the interaction of eIF-2 with the recycling factor GEF (guanine nucleotide exchange factor) [29]. A final implication of these findings is that the primary structure of eIF-2/~ starts with the first methionine in the cDNA-sequence obtained by Hershey and co-workers [15], rather than at one of the other methionines near the 5' end of its predicted amino acid sequence (at positions 6 and 12). Acknowledgements We are very grateful to Dr. John Hershey and his colleagues at the University of California (Davis) for allowing us to quote their data on the sequence of eIF-2/~ prior to full publication. We wish to thank David Colthurst for his excellent assistance and in particular for preparing the casein kinase-2 used in this work, and Dr. David Campbell (University of ~'undee) for performing the amino acid analysis. This work was supported by a grant (to C.G.P.) from the Science and Engineering Research Council, U.K. References 10choa, S. (1983) Arch. Biochem. Biophys. 223, 325-349. 2 Proud, C.G. (1986) Trends Biochem. Sci. 11, 73-77.
3 Pain, V.M. (1986) Biochem. J. 235, 625-637. 4 Issinger, O.G., Benne, R., Hershey, J.W.B. and Traut, R.R. (1976) J. Biol. Chem. 251, 6471-6474. 5 Tuazon, P.T., Merrick, W.C. and Traugh, J.A. (1980) J. Biol. Chem. 255, 10954-10958. 6 DePaoli-Roach, A.A., Roach, P.J., Pham, K., Kramer, G. and Hardesty, B. (1981) J. Biol. Chem. 256, 8871-8874. 7 Schatzmann, R.C., Grifo, J.A., merrick, W.C. and Kuo, J.F. (1983) FEBS Lett. 159, 167-170. 8 Clark, S.J., Colthurst, D.R. and Proud, C.G. (1988) Biochim. Biophys. Acta 968, 211-219. 9 Duncan, R. and Hershey, J.W.B. (1984) J. Biol. Chem. 259, 11882-11889. 10 Duncan, R. and Hershey, J.W.B. (1985) J. Biol. Chem. 260, 8493-8497. 11 Colthurst, D.R., Campbell, D.G. and Proud, C.G. (1987) Eur. J. Biochem. 166, 357-363. 12 Colthurst, D.R. and Proud, C.G. (1986) Biochim. Biophys. Acta 868, 77-96. 13 Tavar6, J.M. and Denton, R.M. (1988) Biochem. J. 252, 607-615. 14 Holmes, C.F.B., Campbell, D.G., Caudwell, F.B., Aitken, A. and Cohen, P. (1986) Eur. J. Biochem. 155, 173-182. 15 Ernst, H., Pathak, V. and Hershey, 3.W.B. (1986) Fed. Proc. 45, 1768. 16 Wettenhail, R.E.H., Kudlicki, W., Kramer, G. and Hardesty, B. (1986) I. Biol. Chem. 261, 12444-12447. 17 Hemmings, B.A., Aitken, A., Cohen, P., Rymond, M. and Hofmann, F. (1982) Eur. J. Biochem. 127, 473-481. 18 Meggio, F., Marchiori, F., Borin, G., Chessa, G. and Pinna, L.A. (1984) J. Biol. Chem. 259, 14576-14579. 19 Tuazon, P.T., Bingham, E.W. and Traugh, 3.A. (1979) Eur. J. Biochem. 94, 497-504. 20 Kuenzel, E,A. and Krebs, E.G. (1985) Proc. Natl. Acad. Sci. USA 82, 737-741. 21 Schroeder, W.A., Shelton, J.B. and Shelton, J.R. (1969) Arch. Biochem. Biophys. 130, 551-556. 22 Jagus, R., Crouch, D., Konieczny, A. and Safer, B. (1982) Current Topics Cell. Reg. 21, 35-63. 23 Landon, M. (1977) Methods Enzymol. 47, 145-149. 24 Martos, C., Plana, M., Guash, M.D. and Itarte, E. (1985) Biochem. J. 225, 321-326. 25 Schneider, H.R., Reichert, G.H. and lssinger, O.-G. (1986) Eur. J. Biochem. 161, 733-738. 26 Sommercorn, J. and Krebs, E.G. (1987) J. Biol. Chem. 262, 3839-3843. 27 Perez, M., Grande, J. and Itarte, E. (1987) Eur. J. Biochem. 170, 493-498. 28 Sommercorn, J., Milligan, J.A., Lozeman, F.J. and Krebs, E.G. (1987) Proc. Natl. Acad. Sci. USA 84, 8834-8838. 29 Kimball, S.R., Everson, W.V., Myers, L.M. and Jefferson, L.S. (1987) J. Biol. Chem. 262, 2220-2227.
nnA 1O230
377 BBA Report
Casein ldnase-2 phosphorylates serine-2 in the fi-subunit of initiation factor-2 S t e p h a n i e J. Clark, A n t h o n y J. A s h f o r d , N i g e l T. Price a n d C h r i s t o p h e r G. P r o u d Department of Biochemistry, School of Medical Sciences, University of Bristol, Bristol (U.K.)
(Received 10 August1988)
Key words: Caseinkinase-2; Initiationfactor; Proteinphosphorylation;Proteinsynthesis We have previously presented evidence which suggests that casein Hnase-2 phosphofflates a serine residue near the N-terminus of the ~-subunit of the initiation |actor elF-2 (Clark, S.J. et al. Biochim. Biophys. Acta 968, 2 | | - 2 1 9 ) . We now report further data which confirm that it is serine-2 which is phosphory|ated by casein kinase-2. This data includes (1) the electrophoretic mobi|ities of the phosphopepfides produced by different cleavage techniques, (2) the amino acid composition of the principal phosphopepfide generated by treatment with cyanogen bromide and (3) the resistance of this phosphopepfide to Edman degradation.
Initiation factor-2 (elF-2) mediates the binding of the initiator methionyl-tRNA (Met-tRNAi) to the 40 S ribosome subunit during peptide chain initiation ia eukaryotic cells, eIF-2 is a heterotrimeric protein composed of subunits termed a (eIF-2a, molecular mass 37 kDa),/~ (eIF-2/~, which migrates anomalously on SDSPAGE: in our system it exhibits an apparent molecular mass of about 54 kDa) and y (eIF-2~,, apparent molecular mass 50 kDa). eIF-2a can be phosphorylated at serine-51 by at least two distinct protein kinases, phosphorylation at this site leads to inhibition of peptidechain initiation by impairing the recycling of elF-2 between successive rounds of initiation (reviewed in Refs. 1-3). eIF-2/t can also be phosphorylated, on serines, by several protein kinases including casein kinase-2 [4-6] and protein kinase C [7]. Phosphorylation by these kinases occurs at different sites which appear to be adjacent in the primary sequence of eIF-2fl [8]. Changes in the level of phosphorylation of eIF-2fl accompany the changes in rates of peptide-chain initiation which result from heat-shock or serum depletion in HeLa cells [9,~0]. Here we present data showing that serine-2 of eIF-2fl is phosphorylated by casein kinase-2. Chemicals and biochemicals were all obtained from the sources described in ref. 8. The proteinase from
Abbreviations" elF-2, eukaryoticinitiationfactor-2; elF-2a, -2/~ and -2y, a-, /I- and ~,-subunits, respectively, of eukaryotic initiation factor-2. Correspondence: C.G. Proud, Departmentof Biochemistry,Schoolof Medical Sciences, Universityof Bristol, Bristol BS8 1TD, U.K.
Pseudomonas fragi which cleaves peptide bonds on the C-terminal side of aspartic acid residues was a kind gift from Dr. G.R. Drapeau, University of Montreal. Casein kinase-2 was prepared from rat liver and assayed as described previously [8]. elF-2 was isolated from rabbit reticulocytes as described in Ref. 11. elF-2 was phosphorylated by casein kinase-2 (0.5 U per #g elF-2) at 30 °C for 40 rain in reaction mixtures containing 3-~Nmorpholino)propane sulphonic acid (50 mM, pH 7.0), MgCI 2 (1 mM), KCI (150 raM), fl-mercaptoethanol (7.5 raM) and [y-32p]ATP (0.1 mM spec. radioact. 0.5 Ci. retool-t). Reactions were terminated by boiling in SDS sample buffer prior to loading onto SDS-polyacrylamide gels. Gels contained 12.5~ (w/v) acrylamide and 0.13~ (w/v) methylene bis-acrylamide as described previously [12] and, under these conditions, eIF-2fl is the slowest migrating subunit. After fixing, staining and destaining the gel, the band corresponding to eIF-2fl was excised and the fl-subunit was electroeluted in buffer containing 30 mM Tris-HC! (pH 7.8), 2 mM EDTA, 0.05~ (w/v) SDS and 15 mM fl-mercaptoethanol [13]. The eluate was dialysed for 6 h against 50 mM ammonium hydrogen carbonate, pH 7.8, and then concentrated in a Speed Va¢ vacuum concentrator (Savant, Hicksville, NY, USA). Protein was precipitated by adding acetone (4 volumes) at - 8 0 ° C for 1 h. The precipitated protein was then dissolved in either 50 mM ammonium hydrogen carbonate (pH 7.8) (for proteinase digestion) or 70~ (v/v) formic acid (for cyanogen bromide digestion). Aliquots (50 pl) of redissolved peptide in 50 mM ammonium hydrogen carbonate (pH 7.8) were treated
0167-4889/89/$03.50 © 1989ElsevierSciencePubfishersB.V.(BiomedicalDivision)
378 with Staphylococcus aureus B8 proteinase (10 ~g. m1-1) or P. fragi proteinase (10/tg. ml -]) for 2 h at 37°C, followed by addition of an equal amount of fresh proteinase and incubation was continued for a further 18 h. Other samples were digested with cyanogen bromide (100 mg. ml -~) in a final concentration of 70~o (v/v) formic acid [8]. After digestion, samples were dried in the vacuum concentrator and then dissolved either in 0.1~ (v/v) trifluoroacetic acid (for reverse-phase chromatography) or in 50~ (v/v) formic acid (for application onto thin-layer plates) [8]. Amino acid analysis was performed for us by Dr. David Campbell at the Department of Biochemistry, University of Dundee. After hydrolysis for 16 h at I I 0 ° C in 6 M HCI/2 mM phenol, samples were analysed using a Waters Picotag system as described in Ref. 14. The N-terminal sequence of eIF-2/~, obtained by eDNA cloning techniques [15] is shown below (Scheme I) (X indicates that the N-terminus is blocked [16], serines are shown bold and acidic residues are underlined). From the complete sequence of elF-2fl and the known substrate specificities of casein kinase-2 and protein kinase C, together with data showing that the sites in elF-2~ which are labelled by casein kinase-2 and protein kinase C are located in a single tryptic peptide, we have previously deduced that the phosphorylation sites are situated near the N-terminus of elF-2~ [8]. This tryptic l~eptide probably represents residues 1-14 of elF-2/~, and was previously reported to contain all the label intr,~aced by casein kinase-2. Within this peptide, ~erine-2 has adjacent to it, on the C-terminal side, two ~c~dic residues, Asp-4 and Glu-5: C-terminal acidic resm~.~ are the substrate specificity determinant for casein kinase-2 [17-20]. The N-terminal sequeace thus contains potential cleavage sites for S. aureus V8 proteinase (glutamic acid residues), P. fragi proteinase (aspartic acid residues) and also cyanogen bromide (methioni_ne residues). Treatment of elF-2/3 labelled by casein kinase-2 with cyanogen bromide, S. aureus V8 proteinase or P. fragi proteinase gave, in each case, a single anionic phosphopeptide (Fig. 1). Given that the N-terminus of elF-2/~ is apparently blocked [16], their electrophoretic properties are entirely consistent with their being derived from the N-terminal sequence of elF-2/~, i.e., X-Met-Ser(P)Gly-Asp-Glu-Met (where X represents the unknown blocking group). The peptides generated by each proce-
5
Q
Q
6 Fig. 1. Elec~rophoretic analysis at pH 3,6 of phosphopeptides derived from elF-2/3 labelled by casein kinase-2. The /3-subunit of elF-2 phosphorylated by casein kinase-2 was subjected to digestion by cyanogen bromide (lane 1), P. fragi proteinase (lane 2) or S. aureus V8 proteinase (lane 3) and the resulting peptides were analysed by electrophoresis at pH 3.6. The figure is an autoradiograph of the dried electrophoresis plate: the polarity ( + / - ) and the origin (arrow) are indicated. The square in lane 1 indicates the migration of the marker dinitrophenylaspartate.
dure would then be X-Met-Ser(P)-Gly-Asp-Glu-homoserine (cyanogen bromide cleavage product), X-MetSer(P-Gly-Asp-Glu (from V8 digestion) and X-MetSer(P)-Gly (generated by the P. fragi proteinase). The Met-Ser bond is clearly a possible additional site for cleavage by cyanogen bromide. However, the data shown in Fig. 1 (relative electrophoretic mobifities) and the amino acid composition data presented below show that cleavage did not occur at this bond. It is known that cyanogen bromide cleaves poorly at Met-Ser bonds [21]. On two occasions, a minor (less than 5% of total) radioactive species was observed migrating faster than the main cyanogen bromide digestion product (not shown): this may represent the X-Met-Ser(P) species. It did not represent another phosphorylation site located elsewhere in the sequence, since digestion of the labelled elF-2p with V8 or trypsin again only gave a single radiolabelled species (not shown, see also below). Based
10
15
X-Met-Sm'-Gly-Asl)-Glu-Met-Ile-Phe.As_.._~Pro.Th r.Met.Ser.Lys.Lys.Lys.Lys.Lys.Lys. Scheme I
379 Peok 20
1,¢
N
---
0 0 0 20 40 60 80 Fig. 2. Reverse-phase analysis of phosphopeptides generated by cyanogen bromide-digestion of elF-2fl phosphorylated by casein kinase-2. 32p radioactivity (o o); acetonitrile concentration (~). Horizontal bars labelled 1, 2 and 3 indicate the corresponding pools of radioactive material referred to in the text.
on the observation that the different cleavage procedures yidded peptides differing by only one or two amino acids in size but with markedly different electrophoretic mobilities (Fig. 1), the blocking group would appear to be small (of the order of the mass of an amino acid) and not highly charged. From the foregoing, it therefore seems that there is only a single major site for phosphorylation by casein kinase-2 in eIF-2fl, at least under our experimental conditions. Other workers have reported incorporation of up to two mol of phosphate per mall of enzymically dephosphorylated elF-2fl [5,22], but ~.~'.' have never observed stoichiometries greater than ~ mol of phosphate per mole of eIF-2fl. However, iF Jur more recent experiments, some of the label (usuahy less than 5%) remained at the origin during electrophoresis (Fig. 1). This differs from our earlier report, and may represent an additional, relatively minor, site for phosphorylation by casein kinase-2. This variability, in our hands, may
TABLE !
Amino acid composition of major phosphopeptidefrom elF-2fl labelled by casein kinase-2 All quantities have been corrected by subtraction of the values obtained when a similar volume of reverse-phase column buffer was analysed. Based on the associated radioactivity, approx. 130 pmol of phosphopeptide were analysed. Amino acid
Amount (pmol)
Aspartic acid Glutamic acid Serine Glycine Homoserine Methionine Alanine Valine All others
79 128 140 152 72 42 19 21 <8
be due to differing endogenous phosphate content in our purified eIF-2 preparations. The peptides resulting from cyanogen bromide cleavage of eIF-2fl labelled by casein kinase-2 were purified by reverse-phase chromatography on a C18 column [8]. The major radioactive peak e:ated at about 3% (v/v) acetonitrile (Fig. 2). Amino acid analysis of the peak fraction gave the composition shown in Table I. This amino acid composition of the phosphopeptide from eIF-2fl labelled by case:n kinase-2 is entirely consistent with phosphorylation at serine-2. The homoserine arises from the cyanogen bromide cleavage, whilst the methionine is probably present for reasons outlined above (i.e., it represents the N-terminal methionine). No precautions (other than carrying out the acid hydrolysis in vacuo) were taken to prevent oxidation of the methionine so the relatively poor recovery of this amino acid is not surprising. No sequence in the primary structure of elF-2fl, other than that at the N-terminus, could give rise to the composition obtained here [15]. On reverse-phase chromatography of the cyanogen bromide digest, two minor phosphopeptides were sometimes observed (Fig. 2). Together they accounted for 10-15% of the total radioactivity: one (peak 1) eluted slightly earlier from the C~8 column and the other eluted at about 5% (v/v) acetonitrile (peak 3). They may represent the products of cyanogen bromide cleavage at the first Met-Ser(P) bond giving Ser(P)-GlyAsp-Glu-Hser and at the Asp-Pro bond giving X-MetSer(P)-Gly-Asp-Glu-Met-Ile-Phe-Asp. Asp-Pro bonds are susceptible to hydrolysis in high concentrations of formic acid such as those (70% (v/v)) in which cyanogen bromide digestion is carried out [23]. The peptides in peaks 1 and 3 did not represent other phosphorylation sites in elF-2fl, since digestion of the same sample of phosphorylated elF-2/] by trypsin, V8-proteinase or P. fragi proteinase gave in each case only a single phosphopeptide species on electrophoresis or peptide mapping (not shown, but see Ref. 8). The major cyanogen bromide-generated phosphopeptide (peak 2) was refractory to the Edman degradation (result not shown). This observation indicates that it does indeed represent the N-terminal peptide of eIF-2fl, whose N-terminus is known to be blocked [16] and thus confirms that cyanogen bromide does not remove the N-terminal methionine from the major phosphopeptide species. As discussed above, casein kinase-2 phosphorylates serine or threonine residues which have several acidic amino acid residues immediately adjacent on the Cterminal side [17-20]. Casein kinase-2 generally 'prefers' substrates with larger clusters of acidic residues than the two adjacent to serine-2 in elF-2fl (Asp-4, Glu-5), but eIF-2fl does contain a glutamic acid residue at the particularly important +3 position [18]. Since serine-2 is the site phosphorylated by casein kinase-2
380
and since the casein kinase-2 and protein kinase C sites are located in the same tryptic peptide [8], the residue phosphorylated by protein kinase-C must be serine-13 as predicted from the data presented in Ref. 8. The activity of casein kinase-2 is known to undergo changes under a variety of physiological conditions and appears, for example, to be activated in response to insulin and epidermal growth factor [24-28]. It will, therefore, be of interest to discover whether the alterations in the level of phosphorylation of eIF-2~ which have been observed in response to various stimuli in HeLa cells [9,10] are occurring at the site phosphorylated by casein kinase-2. Having now identified the site labelled by casein kinase-2, it should be possible to investigate this. Although it is unclear how, if at all, phosphorylation of the/3-subunit affects eIF-2 activity, a recent report has suggested that the/~-subunit may be involved in the interaction of eIF-2 with the recycling factor GEF (guanine nucleotide exchange factor) [29]. A final implication of these findings is that the primary structure of eIF-2/~ starts with the first methionine in the cDNA-sequence obtained by Hershey and co-workers [15], rather than at one of the other methionines near the 5' end of its predicted amino acid sequence (at positions 6 and 12). Acknowledgements We are very grateful to Dr. John Hershey and his colleagues at the University of California (Davis) for allowing us to quote their data on the sequence of eIF-2/~ prior to full publication. We wish to thank David Colthurst for his excellent assistance and in particular for preparing the casein kinase-2 used in this work, and Dr. David Campbell (University of ~'undee) for performing the amino acid analysis. This work was supported by a grant (to C.G.P.) from the Science and Engineering Research Council, U.K. References 10choa, S. (1983) Arch. Biochem. Biophys. 223, 325-349. 2 Proud, C.G. (1986) Trends Biochem. Sci. 11, 73-77.
3 Pain, V.M. (1986) Biochem. J. 235, 625-637. 4 Issinger, O.G., Benne, R., Hershey, J.W.B. and Traut, R.R. (1976) J. Biol. Chem. 251, 6471-6474. 5 Tuazon, P.T., Merrick, W.C. and Traugh, J.A. (1980) J. Biol. Chem. 255, 10954-10958. 6 DePaoli-Roach, A.A., Roach, P.J., Pham, K., Kramer, G. and Hardesty, B. (1981) J. Biol. Chem. 256, 8871-8874. 7 Schatzmann, R.C., Grifo, J.A., merrick, W.C. and Kuo, J.F. (1983) FEBS Lett. 159, 167-170. 8 Clark, S.J., Colthurst, D.R. and Proud, C.G. (1988) Biochim. Biophys. Acta 968, 211-219. 9 Duncan, R. and Hershey, J.W.B. (1984) J. Biol. Chem. 259, 11882-11889. 10 Duncan, R. and Hershey, J.W.B. (1985) J. Biol. Chem. 260, 8493-8497. 11 Colthurst, D.R., Campbell, D.G. and Proud, C.G. (1987) Eur. J. Biochem. 166, 357-363. 12 Colthurst, D.R. and Proud, C.G. (1986) Biochim. Biophys. Acta 868, 77-96. 13 Tavar6, J.M. and Denton, R.M. (1988) Biochem. J. 252, 607-615. 14 Holmes, C.F.B., Campbell, D.G., Caudwell, F.B., Aitken, A. and Cohen, P. (1986) Eur. J. Biochem. 155, 173-182. 15 Ernst, H., Pathak, V. and Hershey, 3.W.B. (1986) Fed. Proc. 45, 1768. 16 Wettenhail, R.E.H., Kudlicki, W., Kramer, G. and Hardesty, B. (1986) I. Biol. Chem. 261, 12444-12447. 17 Hemmings, B.A., Aitken, A., Cohen, P., Rymond, M. and Hofmann, F. (1982) Eur. J. Biochem. 127, 473-481. 18 Meggio, F., Marchiori, F., Borin, G., Chessa, G. and Pinna, L.A. (1984) J. Biol. Chem. 259, 14576-14579. 19 Tuazon, P.T., Bingham, E.W. and Traugh, 3.A. (1979) Eur. J. Biochem. 94, 497-504. 20 Kuenzel, E,A. and Krebs, E.G. (1985) Proc. Natl. Acad. Sci. USA 82, 737-741. 21 Schroeder, W.A., Shelton, J.B. and Shelton, J.R. (1969) Arch. Biochem. Biophys. 130, 551-556. 22 Jagus, R., Crouch, D., Konieczny, A. and Safer, B. (1982) Current Topics Cell. Reg. 21, 35-63. 23 Landon, M. (1977) Methods Enzymol. 47, 145-149. 24 Martos, C., Plana, M., Guash, M.D. and Itarte, E. (1985) Biochem. J. 225, 321-326. 25 Schneider, H.R., Reichert, G.H. and lssinger, O.-G. (1986) Eur. J. Biochem. 161, 733-738. 26 Sommercorn, J. and Krebs, E.G. (1987) J. Biol. Chem. 262, 3839-3843. 27 Perez, M., Grande, J. and Itarte, E. (1987) Eur. J. Biochem. 170, 493-498. 28 Sommercorn, J., Milligan, J.A., Lozeman, F.J. and Krebs, E.G. (1987) Proc. Natl. Acad. Sci. USA 84, 8834-8838. 29 Kimball, S.R., Everson, W.V., Myers, L.M. and Jefferson, L.S. (1987) J. Biol. Chem. 262, 2220-2227.