Nov 24, 1992 - adenosine, 1 mM uridine, 1 mM alanosine, and 0.05 mM 2'-deoxyco- formycin (22). Clones were subsequently selected stepwise for increas-.
THEJOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc
Vol. 268, No. 17, Issue of June 15, pp. 1294612951,1993 Printed in U.S.A.
Expression of a Phosphorylation-resistant Eukaryotic Initiation Factor 2 a-Subunit Mitigates Heat Shock Inhibitionof Protein Synthesis* (Received for publication, November 24, 1992, and in revised form, January 15, 1993)
Patricia Murtha-Riel$,Monique V. Davies$, BradleyJ. Scherers, Sang-Yun Choitll, John W. B. Hershey#, and Randal J. KaufmanSII From the $Genetics Institute, Cambridge, Massachusetts 02140 and the §Department of Biological Chemistry, University of California, Davis, California95616
Protein synthesis is dramatically reduced upon ex- to phosphorylation on thea-subunit (4). When eIF-Pa is posure of cells to elevated temperature. Concordant phosphorylated, the eIF-2. GDP. eIF-2B complex is stabilized with this inhibition, multiple phosphorylation and de- and GTP exchange cannot occur (5-7). Since small increases phosphorylation reactions occur on specific eukaryotic in the degree of eIF-2a phosphorylation are associated with initiation factors thatare required for proteinsynthe- translational inhibition, it is proposed that inhibition occurs sis. Most notably, phosphorylation of the a-subunit of as a resultof sequestering limiting amounts of eIF-2B by the eukaryotic initiation factor-2 (eIF-2a) on serine resi- phosphorylated eIF-2. GDP complex. due 51 occurs. To identifytheimportance ofphosTwo highly specific eIF-2a kinases have been extensively phorylation in control of protein synthesis, we have characterized. The hemin-controlled kinase (HCR) was first evaluated theeffects of expressionof a mutant eIF-2a characterized in reticulocyte lysate deprived of hemin (8, 9). which is resistant to phosphorylation. Expression ofa heat-shocked cells occurs serine to alanine mutant at residue 51 of eIF-2a par- Phosphorylation of eIF-2ain through activation of an HCR-like eIF-2a kinase (10). The tially protected cells fromtheinhibitionofprotein synthesis in response to heat treatment. The overex- double-stranded RNA-activated protein kinase (DAI or repressed serine to alanine5 1 mutant subunitwas incor- cently termed PKR) is induced by interferon, and itsactivity porated into theeIF-2 heterotrimer andwas resistant is dependent on double-stranded RNA (11). Upon viral infecto phosphorylation. These results are consistent with tion, induction and activation of DAI kinase occurs as partof the hypothesis that heat shock inhibition of translation the host antiviral response and results in a generalized inhiis mediated in part through phosphorylationeIF-2a. of bition of protein synthesis. Expression of the wild type or mutant eIF-2a did not Both HCR and DAI kinases phosphorylate a common affect cell survival or induction ofhsp7O mRNA upon serine in eIF-2a, specifically residue 51 (12). Although it was heat shock, indicating that although eIF-2a is a heat originally proposed that serine 48 was apotentialsite of shock-induced protein, its increased synthesis during phosphorylation (13), more recent data suggests that it is heat shock does not alter the heat-shock response. unlikely that phosphorylation at a second site occurs (14-16). Previously, the importance of eIF-2a phosphorylation in translational control in uiuo was demonstrated by expression of eIF-2a wild type and serine to alanine mutants at both Protein synthesis initiates when the ternary complex of putative sites of phosphorylation, residues 48 and 51 (15-17). eIF-2,’ GTP, and Met tRNA binds 40the S ribosomal subunit. Expression of either mutant, but not wild type, protected Following the binding of mRNA, the 60 S ribosomal subunit adenoviral mRNAs and plasmid-derived mRNAs from inhijoins the 48 S preinitiation complex, with concomitant hy- bition of translation mediated by DAI kinase. Whereas the drolysis of GTP and release of eIF-2 bound to GDP (1). For alanine 51 mutant was not phosphorylated, the alanine 48 eIF-2 to promote another round of initiation, the GDP must mutant was phosphorylated. Furthermore, expression of a be exchanged for GTP, a reaction catalyzed by the guanine serine 51 to asparticacid mutant, createdto mimic the charge nucleotide exchange factor (eIF-2B) (2,3). eIF-2 consists of a of a phosphorylated serine, inhibitedprotein synthesis, heterotrimer of three nonidentical subunits: a (36 kDa), p (38 whereas expression of a serine 48 to aspartic acid mutant had kDa), andy (52 kDa). Regulation of eIF-2 utilizationis subject no effect (14, 15).Additional mutation of the serine 48 to an * This work was supported inpart by National Institutes of Health alanine within the aspartic acid 51 mutant partially restored Grant GM22135 (to J. W. B. H.) and National Institutes of Health protein synthesis, showing that the serine hydroxyl group at Training Grant GM07577 (to B. J. S.). The costs of publication of residue 48 is required for efficient inhibition mediated by this article were defrayed in part by the payment of page charges. phosphorylation at residue 51 (14). Taken together, these This article must therefore be hereby marked “aduertisement” in results demonstrated that eIF-2a phosphorylation at residue accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1Present address: Dana Farber Cancer Center, Harvard Medical 51 is an essential control mechanism for initiation of protein synthesis under conditions of DAI kinase activation. School, Boston, MA 02115. Upon heat treatment of cells, eIF-2a becomes phosphoryl(1 To whom correspondence should be addressed. Tel.: 617-8761170; Fax: 617-876-1504. atedandproteinsynthesis is inhibited (18-20). However, The abbreviations used are: eIF-Za, a-subunit of eukaryotic ini- there is not a strict correlation between inhibition of protein tiation factor 2; PAGE, polyacrylamide gel electrophoresis; ADA, adenosine deaminase; CHO, Chinese hamster ovary; HCR, hemin- synthesis in response to heat shock and phosphorylation of controlled kinase; DAI, double-stranded RNA-activated protein ki- eIF-2a. Inhibition of protein synthesis in response to mild heat treatment canoccur in the absence of significant eIF-2a nase.
12946
eIF-2a Phosphorylation Inhibits Protein Synthesis upon Heat Shock
12947
phosphorylation (19). Therefore, there is uncertainty regard- loaded onto a Pharmacia HR 10/30 Superose 12 fast protein liquid ing the significance of eIF-2a phosphorylation for inhibition chromatography column. The column was washedand eluted with 20 mM HEPES, pH 7.5,500 mM KCI, 2 mM EDTA, 50 mM NaF, 7 mM of protein synthesis inresponse to heat. To assess the impor-2-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride, 10%glyctance of eIF-2a phosphorylation in heat shock-mediated in- erol. Fractions were collected and supplementedwith 100 mg ofcarrier hibition of translation, Chinese hamster ovary (CHO) cells lysozyme and quicklyfrozen and stored a t -70 "C. The fractions were derived which overexpress the wild type, the Ala-48, or corresponding to free eIF-2a and eIF-2 complex were identified by peak the Ala-51 mutant eIF-2amolecule by selection for co-ampli- immunoblotting with affinitypurified antibodies to eIF-2a. The fication of cDNA expression vectors with a n adenosine de- fractions corresponding to eIF-2 complex were analyzed hy vertical isoelectric focusingPAGE and immunoblotting using rabbit antiaminase (ADA) gene. In this reportwe have studied theeffect slab human eIF-2a antiserum as described (29). Band intensities were of wild type and mutant eIF-2a overexpression on the inhi- quantitated using an LKB UltroScan XL laser densitometer (Pharbition of protein synthesis mediatedby heat shock. macia Inc., Uppsala, Sweden). Control experimentswere performed on HeLa cells and CHOcells before and after a 45-min heat treatment a t 44.5 "C. Cell extracts were prepared and fractionated by chromatography as above. Equal Derivation of Wild Type and eIF-2a MutantOverexpressing CellsThe mutagenesis of serine 48 or serine 51 to alanine was described amounts of eluted protein from the peak fraction containing heteropreviously (15). The eIF-2a wild type or mutant cDNAs encoded trimeric complex eIF-2 were analyzed by vertical slab isoelectric within a 1.6-kilobase EcoRI fragment were cloned into the cDNA focusing PAGE and immunoblotting using a monoclonal antibody expression vector pMT3SV2ADA (21). eIF-2a mRNA is transcribed directed against eIF-2a (mAb 167.6.23, kindly provided by R. Panfrom the adenovirus major late promoter with SV40 enhancer ele- niers)(20) anddeveloping with anti-mouse IgG-alkaline phosphatase conjugate (Organon Teknika, Durham, NC)using chemiluminescent ment. Thisvector also contains an adenosine deaminase transcription unit under controlof the SV40 early promoter for selection of recip- substrate (Boehringer Mannheim). ient cells for resistance to high concentrations of adenosine in the RESULTS presence of the adenosine deaminase inhibitor 2'-deoxycoformycin (22). DNA was precipitated withcalcium phosphate and transfected Chinese hamster ovary cells were derived that overexpress into CHO-M cells as described (23). CHO-M cells were previously either eIF-Pa wild type, or serine to alanine mutations at derived by co-transfection of colony stimulating factor-1 (M-CSF) residues 48 or 51by coamplification with adenosinedeaminase and dihydrofolate reductaseintodihydrofolate reductase-deficient CHO cells (23, 24). Selection for ADA was accomplished by subcul- (22). The steady state levels of eIF-2a in the overproducing turing cells a t 48 h post-transfection in medium containing 1.1 mM cell lines were evaluated by Western blot analysis using an adenosine, 1 mM uridine, 1 mM alanosine, and 0.05 mM 2'-deoxycoanti-eIF-Pa polyclonal antibody (Fig. 1).Compared to a cell formycin (22).Clones were subsequently selected stepwise for increas- line transfected with the adenosine deaminase vector alone ing resistance upto 1 mM 2'-deoxycoformycin. Characterization ojeZF-2a Expression-To monitor eIF-2a expres- and selectedfor adenosine deaminase expression (lune I), sion, cell extracts were prepared by lysis in Nonidet P-40/SDS lysis expression of the wild type (clone wt(4); lune 2 ) , the Ala-48 buffer (0.15 M NaCI, 0.05 M Tris-HCI, p H 8.0,0.05% SDS, 1% Nonidet mutant (clone 48(2); lune 3), and the Ala-51 mutant (clone P-40, 0.1 mM phenylmethylsulfonyl fluoride) and analyzed by SDS- 51(2); lune 4 ) were elevated approximately 5-fold. A second polyacrylamide gel electrophoresis (PAGE) and Western immunoclone (clone 51(5); lune 5 ) expressedapproximately 3-fold blotting procedures as described (17). eIF-2a synthesiswas analyzed greater levels of Ala-51 eIF-2a compared to control cells. The by pulse labeling cellswith [35S]methionine and immunoprecipitation increased steadystate level of eIF-2acorrelated with inof eIF-2a from cell extracts. Immunoprecipitates were analyzed by creased eIF-Pa mRNAexpression and protein synthesis (data SDS-PAGE and prepared for autoradiography by treatment with ENHANCE (Du Pont-New England Nuclear). Total cellular RNA not shown). Theoverexpression of either wild type or mutant was isolated as described (25) and eIF-2a mRNA was analyzed by eIF-2a subunit did not detectably affect cell growth rate or Northern blot analysis(26). saturation density (data not shown). Heat Treatment of CHO Cells-Conditioned medium was removed The effect of wild type and mutant eIF-Pa overexpression from logarithmically growing cells and medium preheated to 43.5 "C on translation inhibition mediated by heat treatment was was applied. Cellswere incubated by floating on a 43.5 "C water bath "C for 30 min. After monitored after incubating cells a t 43.5 for 30 min. Cells were rinsed with preheated methionine-freemedium this period, protein synthesiswas measured by pulse-labeling and then fed preheated methionine-free medium containing 50 pCi/ ml [3hS]methionine (>lo00 Ci/mmol, Amersham Corp.). After incu- cells with ["S]methionine for 15 min at theelevated temperbation for an additional 15 min at the elevated temperature, cells ature. Incorporation of [:"S]methionine into protein was repwere rinsed in cold phosphate-buffered saline and cell lysates preresented as apercentage of incorporation into non-heated pared in Nonidet P-40/SDSlysis buffer. Protein concentrations were determined by the Bradford protein assay (27). Equal amountsof cell protein were analyzed by reducing SDS-PAGE. Autoradiographywas o-oo" performed after treatment with ENHANCE. To monitor induction u 3 " M M of hsp70 mRNA, total cellular RNA from cells treated for 2 h a t 43.5 "C was isolated (25) and analyzedby Northern blothybridization 92analysis (26) using a human hsp7O probe (28) prepared by random priming with oligonucleotidesas described by the supplier (Pharmacia LKB Biotechnology, Inc.). 48 The effect of heat treatment onsurvival was monitored by plating ceIF-2a control cells and cells heat-treated for 45 min at 43.5 "C a t 100 or 1000 cells/6O-mm plate in triplicate. After10 days, the colonies were 30 stained with methylene blue and counted. The values presented are the average of the triplicate plates which generally variedby less than I 2 3 4 5 10%. and Heat-shocked CellsAnalysis ojeIF-2 Heterotrimers Control in FIG. 1. Overexpression of wild type and serine 51 and 48 Control and 30-min43.5 "C heat-treated cells (5 X lo6) were washed to alanine mutants of eIF-Sa in CHO cells. Western immunoblot with cold phosphate-buffered saline and extracts preparedby lysis in analysis of eIF-2a overexpressing cells using rabbit anti-human eIF20 mM Tris-HCI, pH 7.4,50 mM KCI, 2 mM MgC12,50 mM NaF, 2.0% 2a wasperformed as described under"Materialsand Methods." Nonidet P-40, 1% aprotinin, 1 mM phenylmethylsulfonyl fluoride, 1 Clones shown were isolated from transfection with ADA vector alone mg/ml soybean trypsin inhibitor, 20 mg/ml pepstatin, and 20 mg/ml (ADA, lane 1 ); eIF-2a wild type vector (wt(4), lane 2); eIF-2a Ala-48 leupeptin. The cell lysate was centrifuged at 5,000 X g for 10 min. mutant (48(2), lane3 ) ;and eIF-2a Ala-51 mutants (51(2),lane 4, and The supernatant (cytosolic extract) was brought to 500 mM KC1 and 51(5), lane 5). MATERIALS AND METHODS
aEci35 -
I
12948
eIF-2a Phosphorylation Inhibits Protein Synthesis
cells in six independent experiments (Table I). Heat treatment of the control cell line ADA or the wild type eIF-2a overexpressing cell line wt(4) reduced protein synthesis to values between 4 and 25% of control. In cells expressing either mutant eIF-2a, protein synthesis rates ranged from 3- to 5fold higher than control heat-treated cells. Although the extent of heat shock inhibitioninproteinsynthesis varied between the different experiments on different days, similar trends were observed within any one experiment. These results suggest that eIF-Pa phosphorylation contributes to the inhibition of protein synthesis under conditions of moderate (43.5 "C) heat shock. As the severity of the heat shock increased with increasing temperature, there was also a significant reduction in protein synthesis in the 51(2) cells (data not shown), indicating that other mechanisms must be responsible for the inhibition observed under conditionsof more severe heat shock (19). To determine whetherthe expression of the mutant eIF-2a specifically rescued asubset of mRNAs from translation inhibition or whether it rescued global protein synthesis, SDSPAGE of pulse-labeled cell extracts was performed. Overexpression of wild type, the Ala-48, or theAla-51 eIF-2cu did not grossly alter thespecificity of proteins synthesized compared to control cells (Fig. 2; compare lanes 3, 5, 7, and 9). In addition, the spectrum of polypeptides synthesized in heattreated cells which express either the Ala-48 or -51 mutants of eIF-Sa was similar to the polypeptides detectably synthesized in control cells (Fig. 2; compare lanes 6 and 8 with a 20fold longer exposure of control cells in lane 11). These results show that the spectrum of polypeptides synthesized in the eIF-Pa mutant expressing cells was not dramatically altered compared to control cells in the presence or absence of heat shock. Expression of the mutant eIF-2a appears to only lessen the severity of the inhibition of global protein synthesis. To elucidate the mechanism by which overexpression of mutant eIF-2a interfered with the heat shock inhibition of protein synthesis, we asked whether the eIF-2a mutantsubunit was phosphorylated and whether the eIF-2a subunit was incorporated into the eIF-2 heterotrimer. For this analysis, eIF-2 heterotrimers from control and heat-treated cells were isolated from excess free eIF-2a by gel-filtration fast protein liquid chromatography and subjected to vertical slab isoelectric focusing gel and Western blot analysis using an anti-eIF2a antibody. First, the migration of nonphosphorylated and phosphorylated forms of human and Chinese hamster eIF-2a were studied by analyzing untreated and heat-treated HeLa
9.2 9.8 9.5 25
upon Heat Shock
-+ -+- + - + -+t
2oo-l h92 69- .
+
0
45-2
HeatTreoted
"- r
-
-
f;
30
I
2 3 4
5
6
7
8
91011
FIG.2. Overexpression of either alanine 48 or 51 mutants of eIF-20 protects from inhibition of global protein synthesis upon heat treatment. SDS-PAGE of pulse-labeled cell extracts before (37 "C,labeled -) and after heat treatment (43.5 "C, labeled +) was carried out as described under "Materials and Methods." The figure shows an autoradiogram of the dried gel. Cell lines analyzed are indicated at the top; molecular weight markers are indicated at the left.
and CHO cells, respectively. For this characterization, cells were heat treated at 44.5 "C in order to elicit significant amounts of eIF-2a phosphorylation. Analysis of control purified human eIF-2a treated in vitro with HCR kinase identified two species (Fig. 3A, lane 1 ) which correspond to nonphosphorylated and phosphorylated human eIF-2a. Analysis of eIF-2heterotrimers isolated from untreatedand heattreated HeLa cells demonstrated an increase in the amount of phosphorylated eIF-Pa from 12 to 20% upon heat treatment (Fig. 3A, compare lanes 2 and 3). The a-subunit in theeIF-2 complex isolated from control CHO cells migrated as a doublet (Fig. 3A, lane 4 ) .The more abundant lower band corresponded to nonphosphorylated eIF-Za, whereas the higher band corresponded to thephosphorylated species. Nonphosphorylated eIF-2a isolated from CHO cells (Fig. 3A, lane 4 ) migrated below nonphosphorylated eIF-2a from HeLa cells, presumably due to differing isoelectric points between human and CHO eIF-2a. Heat treatment of control CHO cells yielded a greater proportion of eIF-2a migrating as the single phosphorylated species, just above nonphosphorylated HeLa eIF2 a (Fig. 3A, lane 5 ) . Next, we analyzed the phosphorylation status of overexpressed wild type and mutant eIF-2a. Analysis of eIF-2a protein in total extractsof the human eIF-20 overexpressing TABLE I CHO cells showed detectable amounts of endogenous eIF-2a Percent ofprotein synthesis in response to heat treatment as well as the endogenous phosphorylated eIF-2a subunit The indicated clones were incubated a t 43.5 "C for 30 min and (Fig. 3B, lanes 3, 5, 7, and 9). In addition, a distinct band then protein synthesis was measured by pulse labeling cells with [35S] (Fig. 3B, lanes 3, 5, 7, and 9) representing the nonphosphormethionine for 15 min at theelevated temperature as described under "Materials and Methods." Cell extracts were prepared and incorpo- ylated human eIF-2cu was also observed in the eIF-Sa expressration into trichloroacetic acid-precipitable protein was measured. ing cells. For the wild type and Ala-48 mutant eIF-2a, there The cpm/pg protein are represented as a percent of the cpmlpg was also a detectablespecies migrating as thephosphorylated protein obtained from a parallel dish labeled a t 37 "C. The results human eIF-2 subunit (Fig. 3B, lanes 3 and 5 ) which was not from six independent experiments are shown. observed for the Ala-51 mutant (Fig. 3B, lanes 7 and 9).Heat Cell line treatment of control transfectedCHO cells (ADA) resulted in Experiment a greater proportion of eIF-Sa migrating as the phosphorylADA wt (4) 48 (2) 51 (2) 51 ( 5 ) ated species (Fig. 3B, lane 2). Upon heat shock of wild type 76 and Ala-48 eIF-2a expressing cells, there was also a relative 1 4.3 287.7 23 33 increase in the more acidic phosphorylated human species 2 16 ND" 9.5 26 compared to the nonphosphorylated human eIF-2a (Fig. 3B, 3 8.1 14 46 30 lanes 4 and 6 ) .Thus, the wild type and Ala-48 mutant were 4 12 9.7 27 38 39 5 10 27 ND 26 detectably phosphorylated in response toheat shock. Al6 21 30 93 ND though the amount of increased phosphorylation was small, the increase observed is consistent with previous observations ND, not determined.
eIF-2a Phosphorylation Inhibits Protein Synthesis
A
HeLa
CHO
” I
HCR
+
+
-
Heat Treated
a A
0
a
I
B
2
5
4
3
wt(4)
ADA
(2)
r”
1
2
c-
3
4
HeLa
5
ADA
5l(S)
+ - +
1 -
6
7
51(2)
8
IO
9
51 (5)
” ”
-
Heat Treated
+ - + - +
Heat Treated
1
2
3 ‘
3
4
5
6
12949
incorporated into the heterotrimer, the eIF-Sa heterotrimer was isolated and analyzed by vertical slab isoelectric focusing gel and Western blotanalysis. In comparison to control CHO cells (ADA)there was a significant amount of a specieswhich co-migrated with the human eIF-2a subunit in heterotrimers isolated from both 51(2) and 51(5) cells (Fig. 3C, lanes 4 and 6). In addition, therewere no detectable species migrating as phosphorylated human eIF-Pa. Thus, the eIF-20 subunit in the eIF-2 heterotrimers from the human eIF-2a Ala-51 mutant transfected cells was primarily composed of the human species and again, no phosphorylated species were detected either before or after heatshock. Since eIF-2a expression is induced upon heat shock (301, we asked whether eIF-Za wild type or mutantoverexpression may alter heat shock gene expression or survival of cells subjected to heatshock. Northern blot analysis demonstrated that inductionof hsp70 mRNA in responseto heat treatment occurred in the presence of mutanteIF-2a (Fig. 4). The slightly lower level of hsp70 mRNA detected incell line 51(5) was likely not a result of eIF-2a mutant expression since cell line 51(2),which expressed higherlevels of the Ala-51 mutant eIF-Ba, exhibited levels of hsp70 mRNA similar to control heat-treated cells. The effect of heat shock on survival was evaluated by monitoring plating efficiency after heat treatment.Expression of wild typeormutanteIF-2aneither improved nor compromised the reduced survival in response to heat shock (Table 11). In two independent experiments where there was a 10-fold (Experiment 1)or a 2-fold (Experiment 2) reduction in survival upon heat treatment, expression of mutant eIF-Pa did not result in significantly altered survival. We concludethat expression of either wild type eIF2a or the mutanteIF-Za which interfered with inhibition of
- Heat
”“8 -.
upon Heat Shock
+ Heat
“
7
FIG. 3. Phosphorylation of wild typeand mutant eIF-2a and incorporation of overexpressed eIF-2a into the eIF-2 complex. Logarithmically growing cells were heat-treated (+) for 30 min (panel A ) , 45 min (panel B ) at 44.5 “C or 30 min at 43.5 “C (panel C) as described under “Materials and Methods” and harvested immediately thereafter. Control cells (-) were harvested similarily, but without heat treatment. PanelsA and C show the eIF-2 heterotrimeric complex isolated by Superose 12 chromatography. Panel B represents analysis of total cytosolic extract. eIF-2a in all samples was analyzed by vertical slab isoelectric focusing PAGE and immunoblotting with anti-eIF-2a antibody as described under “Materials and Methods.” Lane 1 of panel A shows analysis of purified human eIF-2 treated with HCR kinase in uitro prior to analysis. Lane 1 of panel C (labeled HeLa) shows purified human eIF-2 not treated with HCR kinase. Migration of nonphosphorylated endogenous CHO (A), endogenous or transfected human (O), phosphorylated endogenous CHO (A),and endogenous or transfected human (0)eIF-2a are shown. For panel B equal amounts of protein were loaded onto each lane. For panels A and C equal volumes of the fraction containing the eIF-2 heterotrimer were loaded onto each lane. For panel C, a lesser amount of eIF-2 heterotrimer was recovered from heat-treated 51(5) cells (lane 7) as a consequence of cells lost during the heat treatment.
on the effect of heat treatment on eIF-2a phosphorylation (18-20). Upon heat shock of cells expressing the Ala-51 mutant, therewere no detectablemore acidic speciesrepresenting phosphorylatedhumaneIF-2a (Fig. 3B, lanes 9 and 1 1 ) , although a n increase in the amount of endogenous phosphorylated form was observed. These results show that the Ala-51 mutant was not phosphorylated. T o characterize whether the Ala-51 subunit was actually
FIG. 4. Induction of hsp70 mRNA in heat-treated cells. The indicated cell lines were treated for 2 h a t 43.5 “C and total RNA was analyzed by Northern blot hybridization to a human hsp70 probe (28). The asterisk shows a constitutively expressed hsp70 related mRNA.
TABLE I1 Plating efficiency before andafter h a t treatment The indicated clones were incubated a t 43.5 “C for 45 min and plating efficiencies were determined as described under “Materials and Methods.” Efficiencies are presented as the number of colonies over the number of cells plated. Results from twoindependent experiments are presented. Cell line Experiment
Temperature
ADA
w t (4)
“C
1
5.0
7.2 2
26
37 43.5 37 43.5
48 (2)
51 (2)
17
69 4.9 48
%
73 50 35 53
47 43 24 21
24
12950
eIF-2a Phosphorylation Inhibits Protein Synthesis
translation in response to heat, did not interfere with heatinduced gene expression nor affect cell survival in response to heat. DISCUSSION
Mammalian cells rapidly respond to changes in their cellular environment by alteration of their protein synthetic capacity through covalent modification of initiation factors. Increased or decreased protein synthesis in response to different perturbations correlates with phosphorylation and dephosphorylation of multiple initiation factors (4). Prior to inhibition of protein synthesis, eIF-2a becomes phosphorylated in response to heat shock (18-20) and other inducers of the stress response (31-33). In addition, decreased phosphorylation of eIF-4B and eIF-4F correlates with decreased protein synthesis inresponse to heatshock (6,18,34-36) and growth repression (32,37-40). Since multiple phosphorylation and dephosphorylation reactions occur in response to environmental changes, it is difficult to conclude whether the observed effect on the translational apparatus results from any single specific modification. We have approached this question by studying the expression of mutant initiation factors that areresistant to specific phosphorylation reactions to evaluate the importance of any single modification in translational control. The results presented here show that an overexpressed nonphosphorylatable serine to alanine mutant at residue 51 of eIF-2a was incorporated into the majority of eIF-2 heterotrimers to yield an eIF-2 complex that was not phosphorylated upon moderate heat treatment. In the presence of this mutant eIF-2, protein synthesis continuedto a greater extent after a moderate degree of heat treatment. In addition,overexpression of another mutant, Ala-48, which can bypass the effect of phosphorylation at residue 51 (14), also allowed protein synthesis to continue at a greater extent after heat treatment, even in the presence of a phosphorylated serine. The data from these two mutant eIF-2a molecules supports the conclusion that phosphorylation of eIF-2a upon moderate heat shock is a primary mechanism by which translation inhibition occurs. In a detailed characterization of the effect of heat shock on protein synthesis and initiation factor modification, it was concluded that inhibition of protein synthesisdoes not require a change in the modification status of eIF-Pa, eIF-4B,or eIF4E (19). Under mild heat shock (41 and 42 "C), HeLa cell protein synthesis was inhibited significantly, yet little or no phosphorylation of eIF-2a was detected. The result suggests that a mechanism other than eIF-2a phosphorylation is responsible for translational inhibition.At temperatures greater than 42 "C, eIF-2a was phosphorylated, thus representing a second, redundant inhibitorymechanism in these cells. In the experiments reported here, the CHO cell lines were subjected to a moderate (43.5 "C) heat shock, resulting in both translational inhibition and slightly enhanced eIF-2a phosphorylation. The fact that the mutant forms of eIF-2a mitigate translational inhibition indicatesthat the eIF-2a phosphorylation mechanism is a primary cause of repression under these conditions. However, the mutantCHO cell lines are not fully competent for translation at 43.5 "C. Although we cannot rule out thepossibility that theinhibition observed in these heattreated cells results from the residual amount of endogenous phosphorylated eIF-2a inthe eIF-2 complex, it isalso possible that another mechanism of inhibition of protein synthesis is functioning. This other mechanism may be the same mechanism detected in mildly heat-shocked HeLacells. Both studies indicate that multiple, redundant and overlapping pathways
upon Heat Shock
regulate initiation of protein synthesis in heat-shocked cells. These pathways may involve changes in the phosphorylation status of other translational components such as eIF-3, eIF4B, andeIF-4F (35, 36), or to stillother changes not yet identified. Brief exposure of cells to elevated temperature dramatically reduces their survival (41). It is of interest that eIF-Pa has been identified as a gene which is induced upon heat shock (30). We have previously reported that overexpression of eIF2 a in transientlytransfected COS-1 monkey kidney cells induces hsp70 expression (15). Since overexpression of hsp70 renders rat fibroblasts thermotolerant (42), it is possible that increased hsp70 expression may protect cells from inhibition of protein synthesisupon heat shock. However, in the absence of heat shock, overexpression of wild type or mutant eIF-2a in CHO cells did not induce hsp7O mRNA, indicating that increased levels of hsp70 are not responsible for the effect on translation observed. The difference observed in hsp70 induction between the transient DNA transfection experiments in COS-1 cells and the stably transfected CHO cells likely reflects the 10-fold greater levels of eIF-2a expression obtained in the transfected COS-1 cells. Since mutant eIF-2a overexpression permitted protein synthesis under conditions of heatinduced stress, we asked whether the ability to synthesize protein during heat shock might aggravate or improve survival in response to heat treatment. No significant difference was seen in the survival of cells that express mutanteIF-2a compared to control cells. Thus, the inhibition of protein synthesis that occurs during heat shock apparently does not affect cellular survival. 1. 2. 3. 4. 5. 6.
7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
25. 26. 21. 28. 29. 30. 31.
REFERENCES Hershey, J. W. B. (1991) Annu. Rev. Biochem. 6 0 , 717-754 Konieczny, A., and Safer, B. (1983) J. Biol. Chem. 268,3402-3408 Panniers, R., and Henshaw, E. C. (1983) J. Biol. Chem. 2 5 8 , 7928-7934 Hershey, J. W. B. (1989) J. Biol. Chem. 264,20823-20826 Ochoa, S. (1983) Arch. Biochem. Biophys. 223,325-349 Panniers, R., Steward, E. B., Merrick, W. C., and Henshaw, E. C. (1985) J. Bwl. Chem. 260,9648-9653 Safer, B. (1983) Cell 3 3 , 7-8 Matts, R. L., Levin, D. H., and London, I. M. (1983) Proc. Natl. Aead. Sci. U. S. A. 80,2559-2563 Siekierka. J.. Manne. V.. and Ochoa., S. (1984) Proe. Natl. Acad. Sci. U. S. A. . 81,352-356 De Benedetti, A,, and Baglioni, C. (1986) J.Biol. Chem. 261,338-342 Meurs, E., Chong, K., Galabru, J., Thomas, N. S. B., Ken,I. M.,Williams, B. R. G., and Hovanessian, A. G. (1990) Cell 62,379-390 Colthurst, D. R.,Campbell, D. G., and Proud, C. G. (1987) Eur. J. Biochem. 166,357-363 Wettenhall, R. E. H., Kudlicki, W., Kramer, G., and Hardesty, B. (1986) J. Biol. Chem. 261,12444-12447 Choi, S. Y., Scherer, B. J., Schnjer, J., Davies, M. V., Kaufman, R. J., and Hershey, J. W. B. (1992) J. Bml. Chem. 267,286-293 Kaufman, R. J., Davies, M. V., Pathak, V. K., andHershey, J. W. B. (1989) Mol. Cell. Biol. 9, 946-958 Pathak, V. K., Schindler, D., and Hershey, J. W. B. (1988) Mol. Cell. Biol. 8,993-995 Davies, M. V., Furtado, M., Hershey, J. W. B., Thimmappaya, B., and Kaufman, R. J. (1989) Proc. Natl. Acad. Sei. U. S. A. 86,9163-9167 Duncan, R., and Hershey, J. W. B. (1984) J. Biol. Chem. 269,11882-11889 Duncan, R., and Hershey, J. W. B. (1989) J. Cell Biol. 109,1467-1481 Scorsone, K. A., Panniers, R., Rowlands, A. G., and Henshaw, E. C. (1987) J. Biol. Chem 262,14538-14543 Kaufman, R. J. (1990) Methods Enzymol. 186,537-566 Kaufman, R. J., Murtha, P., Ingolia, D. E., Yeung, C.-Y., and Kellems, R. E. (1987) Proc. Natl. Acad. Sci. U. S. A. 83,1568-1571 Kaufman, R. J., Wasley, L. C., Spiliotes, A. J., Gossels, S. D., Latt, S. A,, Larsen, G . R., and Kay, R. M. (1985) Mol. Cell. Biol. 6, 1750-1759 Wong, G. G., Temple, P. A., Leary, A. C., Witek-Giannotti, J. S., Yang, YC., Ciarletta, A. B., Chung, M., Murtha, P., Kriz, R., Kaufman, R. J., Ferenz, C. R., Sibley, B. S., Turner, K. J., Hewick, R. M., Clark, S. C., Yanai, N., Yokota, H., Yamada, M.,Saito, M., Motoyoshi, K., and Takaku. F. (1987) Science 2 3 6 , 1504-1508 Chirgwin, J. M., Przyhyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979) Biochemistry 18,5294-5299 Thomas. P. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,5201-5205 Bradford, M: M. (1976) Anal. Biochem. 72,248-254 Wu, B., Hunt, C., and Morimoto, R. (1985) Mol. Cell. Biol. 6,330-341 Maurides. P. A..Akkaraiu. G. R., and Jams,R. (1989) A d . Biochem. 1 8 3 , 144-151 Colbert, R. A,, Hucul, J. A., Sconone, K. A., and Young, D. A. (1987) J . Biol. Chem. 2 6 2 , 16763-16766 Clemens, M. J., Galpine, A,, Austin, S. A., Panniers, R., Henshaw, E. C., I
1
eIF-2a Phosphorylation Inhibits Protein Synthesis upon Heat Duncan, R., Hershey, J. W. B., and Pollard, J. (1987) J. Biol. Chem. 262,167-111
32. Duncan, R., and Hershey, J. W. B.(1985) J.Biol. Chem. 260,5493-5497 33. Duncan, R.,and Hershey, J. W. B. (1987) MOL. Cell. Bwl. 7,1293-1295 34. Duncan. R..Milburn. S. C.. and Hershev. " . J. W. B. (1987) J. Biol. Chem. 262,3801388
'
35. Lamphear, B. J., and Panniers, R. (1990) J.Bwl. Chem. 266,5333-5336 36. Lamphear, B. J., and Panniers, R. (1991) J. Biol. Chem. 2 6 6 , 2189-2794
Shock
12951
37. Morley, S. J., and Trau h, J. A (1990) J. Biol. Chem. 266,10611-10616 38. Frederlckson R. M. &ntine,'K. S., and Sonenberg, N. (1991) Mol. Cell. Biol. 11,2b96-29h 39. Morley S. J and Traugh J. A. (1989) J. Biol. Chem. 264,2401-2404 40. Morlev: S. J:: Dever. T. E:, Etchison. . . J.Biol. - . . D.. . and Traueh. J. A. (1991) Chem. 266,466914672 41. Li. G. C. and Werb Z. (1982) Proc. Natl. Acad. Sci. U. S. A. 79,3218-3222 42. Li, G. C.'Li L. LidY-KMak J. Y. Chen L., and Lee, W. M. F. (1991) Proc. k t / .A h . h i . U,'S. A,"%, 1k81-1885 '
