Proc. Natl. Acad. Sci. USAVol. 93, pp. 5578-5583, May 1996Biochemistry

Activation of the translational suppressor 4E-BP1 followinginfection with encephalomyocarditis virus and poliovirus

(eIF4F/cap binding protein complex/translational control/picornaviruses)

ANNE-CLAUDE GINGRAS*, YURI SVITKIN*t, GRAHAM J. BELSHAM*t, ARNIM PAUSE*§, NAHUM SONENBERG*¶*Department of Biochemistry and McGill Cancer Centre, McGill University, Montreal, PQ Canada H3G 1Y6; tM. P. Chumakov Institute of Poliomyelitis andViral Encephalitides, Russian Academy of Medical Sciences, Moscow 142782, Russia; and *Biotechnology and Biological Sciences Research Council Institutefor Animal Health, Pirbright, Woking, Surrey GU24 ONF, United Kingdom

Communicated by Richard D. Klausner, National Cancer Institute, Bethesda, MD, January 19, 1996 (received for review June 30, 1995)

ABSTRACT Infection of cells with picornaviruses, suchas poliovirus and encephalomyocarditis virus (EMCV), causesa shutoff of host protein synthesis. The molecular mechanismof the shutoff has been partly elucidated for poliovirus but notfor EMCV. Translation initiation in eukaryotes is facilitatedby the mRNA 5' cap structure to which the multisubunittranslation initiation factor eIF4F binds to promote ribosomebinding. Picornaviruses use a mechanism for the translationof their RNA that is independent of the cap structure. Polio-virus infection engenders the cleavage of the eIF4G (formerlyp220) component of eIF4F and renders this complex inactivefor cap-dependent translation. In contrast, EMCV infectiondoes not result in eIF4G cleavage. Here, we report that bothEMCV and poliovirus activate a translational repressor,4E-BP1, that inhibits cap-dependent translation by binding tothe cap-binding subunit eIF4E. Binding of eIF4E occurs onlyto the underphosphorylated form of 4E-BP1, and this inter-action is highly regulated in cells. We show that 4E-BP1becomes dephosphorylated upon infection with both EMCVand poliovirus. Dephosphorylation of 4E-BP1 temporally co-incides with the shutoff of protein synthesis by EMCV but lagsbehind the shutoff and eIF4G cleavage in poliovirus-infectedcells. Dephosphorylation of 4E-BP1 by specifically inhibitingcap-dependent translation may be the major cause of theshutoff phenomenon in EMCV-infected cells.

Initiation is the rate-limiting step of translation in eukaryotes(1). This step is mediated by the initiation factor eIF4F, whichbinds the cap structure m7GpppX (where X is any nucleotide)that is present at the 5' end of all cellular mRNAs exceptorganellar RNAs (2). eIF4F is comprised of three subunits:eIF4E, eIF4A, and eIF4G (formerly called p220). eIF4E, thecap-binding subunit, is limiting in cells relative to othertranslation factors, thus leading to a restricted pool of eIF4F(3, 4). eIF4F, in combination with a second initiation factor,eIF4B, exhibits RNA helicase activity, which is mediated by theeIF4A subunit (5). Based on this activity, and in light of theevidence that the requirement for the cap structure correlateswith the extent of secondary structure in the mRNA 5'-untranslated region, it was hypothesized that eIF4F unwindsthe secondary structure of the mRNA 5'-untranslated regionto promote ribosome binding (6, 7).

In contrast to cellular mRNAs, picornavirus RNAs are un-capped and initiate translation by internal ribosome binding,which proceeds in a cap- and eIF4E-independent manner (8-10).Infection with most picornaviruses results in the shutoff of hostcell protein synthesis (reviewed in refs. 11 and 12). The mecha-nism by which this occurs is partly understood for some picor-naviruses (enteroviruses, rhinoviruses, and aphthoviruses), but isless clear for cardioviruses, such as encephalomyocarditis virus

(EMCV). The shutoff is exerted at the level of translation,because cellular mRNAs can be recovered from virus-infectedcells in an intact and functionally active form (13-17).

Infection with most members of the picornaviridae family,including enteroviruses (polioviruses and coxsackieviruses),rhinoviruses, and aphthoviruses, results in the cleavage of theeIF4G subunit of eIF4F (reviewed in ref. 12). The modifiedeIF4F is rendered inactive for cap-dependent translation ofcellular RNAs, but remains functional for translation of un-capped viral RNAs by internal ribosome entry (18, 19).However, the cleavage of eIF4G is apparently not the soleevent responsible for the shutoff, because, under certainconditions, complete eIF4G cleavage occurs without totalinhibition of cellular protein synthesis (20, 21).EMCV, a cardiovirus, causes a shutoff of host protein

synthesis, but this inhibition occurs late in infection as com-pared with poliovirus (22). Also, in sharp contrast to poliovi-rus, no cleavage of eIF4G occurs in EMCV-infected cells (23).In vitro translation studies showed a general decrease in theability of EMCV-infected cell extracts to translate both viraland cellular exogenous mRNAs, rather than a specific inhibi-tion of cellular mRNA translation, such as that observed inextracts from poliovirus-infected cells (22, 24-26). This is incontrast to the in vivo studies showing a selective shutoff ofhost protein synthesis.

Recently, a novel protein, 4E-BP1 (eIF4E-binding protein1), which represses the activity of eIF4E, was characterized (27,28). 4E-BP1 is a small heat- and acid-stable protein whoseactivity is regulated by phosphorylation (27-29). The under-phosphorylated form of 4E-BP1 binds to eIF4E and inhibitscap-dependent translation, both in vitro and in vivo (27). Thisinhibition is exerted via the competition of 4E-BP1 and eIF4Gfor the same binding site on eIF4E (30, 31), thus leading to alimited amount of the eIF4F complex. As anticipated, inaccordance with the function of eIF4E in the recognition of thecap structure, 4E-BP1 does not inhibit cap-independent trans-lation (27). Here, we show that 4E-BP1 is activated by de-phosphorylation upon EMCV and poliovirus infections.

MATERIALS AND METHODSVirus Strains. EMCV strains K-2 (26) and Rueckert (R)

(ref. 32; a kind gift from Ann Palmenberg, University ofWisconsin, Madison) were grown in Krebs II cells and HeLaR19 cells, respectively. Mahoney strain of poliovirus type 1 waspropagated in HeLa R19 cells.

Abbreviations: EMC, encephalomyocarditis; EMCV, EMC virus; GST,glutathione S-transferase; CAT, chloramphenicol acetyltransferase.§Present address: Cell Biology and Metabolism Branch, NationalInstitute of Child Health and Human Development, National Insti-tutes of Health, Bethesda, MD 20892.ITo whom reprint requests should be sent at: Department of Bio-chemistry, McGill University, 3655 Drummond Street, Room 807,Montreal, PQ Canada, H3G 1Y6.

5578

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

Proc. Natl. Acad. Sci. USA 93 (1996) 5579

Antibodies. Antibody 11208 was raised in rabbit (PoconoRabbit Farm, Canadensis, PA) against a glutathione S-transferase (GST)-4E-BP1 fusion protein. It does not cross-react with 4E-BP2, a protein that is 56% identical to 4E-BP1(27). The antibody recognizes both human and mouse 4E-BP1.a-eIF4G rabbit polyclonal antibody was kindly provided by LuisCarrasco and Isabel Novoa, Universidad Autonoma de Madrid.

Cell Culture, Infections, and Treatments. Mouse Krebs IIascites carcinoma cells were grown in Balb/c mice for 7 days.Cells were washed twice with Earle's solution (GIBCO) andmock-infected or infected with EMCV at a multiplicity ofinfection of 30 plaque-forming units per cell. After adsorptionof the virus for 30 min at 25°C, mock-infected or EMCV-infected cells (107 cells per ml) were incubated in S-MEM(GIBCO) at 4°C for 14 h. This incubation was performed toincrease the efficiency of infection (33). Cells were thentransferred to 37°C and incubated with gentle agitation insuspension for the times indicated in the figure legends. HeLaR19 cells were grown in DMEM containing 10% fetal bovineserum and infected in serum-free medium at a multiplicity ofinfection of 100 plaque-forming units per cell with EMCV orpoliovirus, respectively.

Metabolic Labeling. Cells were washed and incubated in 1ml of methionine-free DMEM (GIBCO) for 30 min withmethionine (10 ,uCi/ml; 1 Ci = 37 GBq). Cells were lysed inbuffer containing 0.5% Nonidet P-40, 140 mM NaCl, and 30mM Tris HCl (pH 7.5), and nuclei were removed by centrif-ugation. The supernatants from equal numbers of cells wereanalyzed on SDS/12.5% polyacrylamide gels. Radiolabeledproteins were quantified using a Phosphorlmager (model no.Bas 2000; Fuji).Western Analysis. Cells grown in suspension were collected

by centrifugation, and cells grown in monolayers were scrapedin cold buffer A (20 mM Tris, pH 7.5/100 mM KCl/20 mM(3-glycerol phosphate/i mM DTT/0.25 mM Na3VO4/10 mMNaF/1 mM EDTA/1 mM EGTA/lOnM okadaic acid/i mMphenylmethylsulfonyl fluoride). Lysis was performed by threefreeze-thaw cycles, cell debris was pelleted by centrifugation,and the protein concentration in the supernatant was mea-sured by the Bio-Rad assay. To analyze 4E-BP1, 1 mg ofprotein in 500 ,ul was boiled for 7 min and then incubated onice for 10 min. The precipitated material was removed bycentrifugation, and protein in the supernatant was precipitatedwith 15% trichloroacetic acid for 1 h, followed by two washeswith diethyl ether. The pellet was dissolved in Laemmli samplebuffer (34), and aliquots (20%) were loaded on a SDS/15%polyacrylamide gel. Proteins were transferred onto a 0.22-,umnitrocellulose membrane, which was then blocked in 2% milkfor 1 h followed by incubation for 2 h with rabbit polyclonalantiserum 11208 against 4E-BP1 (1:1000) in Tris-bufferedsaline including 0.5% Tween-20 (TBST). For the analysis ofeIF4G, 50 ,ug of total protein extract lysed in cold buffer A(without heat treatment) was loaded on a SDS/8% (or 6%)polyacrylamide gel. Proteins were transferred (5 h at 40 V,without methanol in the transfer buffer) onto nitrocellulosemembrane and analyzed with rabbit anti-eIF4G polyclonalantibody (1:2500 in TBST). Incubation with secondary anti-body was performed with peroxidase-coupled donkey anti-rabbit Ig (1:5000 in TBST). Detection of peroxidase-coupledantibody was performed using chemiluminescence (ECL; Am-ersham). In some experiments, 125I-labeled protein A (Amer-sham) was used (1:1000), and the signal was quantified usingPhosphorlmager analysis.

Far-Western Analysis of4E-BP1. Heat-treated extracts wereprocessed on a SDS/15% polyacrylamide gel as described forthe Western analysis above, and proteins were transferred ontonitrocellulose membranes. Membranes were blocked in 5%milk in HBB (25 mM Hepes-KOH, pH 7.7/25 mM NaCl/SmMMgCl2/1 mM DTT/0.1% Nonidet P-40) for 2 h. Membraneswere then incubated for 12 h in hybridization buffer (20 mM

Hepes-KOH, pH 7.7/75 mM KCl/2.5 mM MgCl2/0.1 mMEDTA/1 mM DTT/0.1% Nonidet P-40/1% skim milk) con-taining 32P-labeled HMK-eIF4E probe at 250,000 cpm/ml, asdescribed (27, 35). After three washes with hybridization buffer,the membranes were dried and exposed against an x-ray film.

In Vitro Translation. Extracts prepared from mock-infectedand EMCV-infected cells were treated with micrococcal nu-clease (36), and translation was performed with [35S]methi-onine in a volume of 12.5 j,l at 30°C for 90 min, as described(37). Following translation, reaction mixtures were supple-mented with 2.5 ,ul of a solution containing 6 mg of RNase A perml and 60 mM EDTA and incubated at 30°C for an additional 5min. Products of in vitro translation reactions were analyzed byelectrophoresis on SDS/15% polyacrylamide gels. Gels wereprocessed for fluorography with En3Hance (DuPont).

In Vivo Labeling and Immunoprecipitation. Monolayers ofHeLa cells were infected with poliovirus as described. At 1 hor 3 h after infection, the medium was removed and cells wererinsed twice in phosphate-free DMEM (GIBCO). [32P]Or-thophosphate (0.5 mCi/ml; DuPont/NEN) was added and cellswere incubated for 2 h with occasional agitation. The medium wasremoved, and cells were rinsed twice in cold TBS and lysed inRIPA buffer (50 mM Tris-HCl, pH 7.5/150 mM NaCl/1%Nonidet P-40/0.5% sodium deoxycholate/0.1% SDS) containing0.25 mM Na3VO4, lOnM okadaic acid, and 10 mM NaF. Debriswas spun down, and extracts were precleared with preimmuneserum preadsorbed on protein A agarose (Repligen) for 1 h at4°C. The supernatant was immunoprecipitated for 4 h at 4°Cusing 11208 antibody preadsorbed on protein A agarose beads.Beads were washed three times in RIPA, once in 200 mM LiCl,1 mM 2-mercaptoethanol, and resuspended in Laemmli samplebuffer (34).

RESULTS

4E-BP1 Phosphorylation Is Reduced in EMCV-InfectedCells. Infection of Krebs II ascites cells with EMCV caused areduction in host cell protein synthesis (Fig. 1A). Shutoff ofhost protein synthesis was detectable 4 h after infection andwas almost complete at 5 h after infection (Fig. 1A, lanes 4-9).Two different strains of EMCV varied in the kinetics of theshutoff. The shutoff was more pronounced for the EMCV K-2than for the EMCV R strain (Fig. 1A, compare lanes 5 and 6).Therefore, the strain EMCV K-2 was chosen for subsequentstudies. To determine whether the 4E-BP1 phosphorylationstate changes upon virus infection, a Western analysis wasperformed with a polyclonal antibody against 4E-BP1. Twoforms of 4E-BP1 were present in mock-infected cells (Fig. 1B,indicated by the two upper arrowheads; the fastest migratingweak band in the gel is probably nonspecific, because it was notdetected by more recent bleeds; lane 1). Between 3 and 4 hafter infection, the intensity of the upper band began todiminish and a faster migrating band appeared (Fig. 1B,indicated by an arrow; lanes 10 and 11). According to previousstudies in mouse cells, the new band is probably the unphos-phorylated form of 4E-BP1, which migrates faster than thephosphorylated forms (38-40). In a recent report, Lin et al.resolved by two-dimensional gel three forms of 4E-BP1. Thefaster migrating form did not contain any phosphate (39). AFar-Western analysis, with an heart muscle kinase-eIF4Eprobe, was performed on the same samples to determinewhether the modification of 4E-BP1 affected its binding toeIF4E. In mock-infected cells, eIF4E interacted only weaklywith the two forms of 4E-BP1 (Fig. 1C, lanes 1-6), againsuggesting that these are the phosphorylated forms, which havea low affinity for eIF4E (27, 28). The signal on the Far-Westernblot with mock-infected cells slightly increased with time (Fig.1C, lanes 1-6), probably as a consequence of serum depriva-tion (A.-C.G., unpublished data). In sharp contrast, when cellswere infected with EMCV, a strong interaction of eIF4E with

Biochemistry: Gingras et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

5580 Biochemistry: Gingras et al.

Atime(h) 3 4 5 6

8 C 8 e 8 r8M A :: ...: a V-' " ..Ws

B mock-infectedtime (h) 1 2 3 4 5 6

21 kDa- _m*

EMCV

1 2 3 4 L6

-.. j

1 2 3 4 5 6 7 8 9 10 11 12p90.4

P.- C

S.. CVPO5F_ <VPI

4_^ -VP3

AAW

7 8 9 10 11 12

mock-infected EMCVtime(h) 1 2 3 4 5 6 1 2 3 4 5 6kDa3-

w1 2 3 4 5 6 7 8 9 10 1112

FIG. 1. EMCV infection leads to 4E-BP1 dephosphorylation. (A) Time course of the shutoff. Krebs II mouse ascites were labeled with[35S]methionine at the indicated times after infection (R and K-2 strains of EMCV), and cytoplasmic protein extracts were analyzed by SDS/PAGE.The arrow indicates an arbitrarily chosen cellular protein (p90) that was used to measure the inhibition of host protein synthesis, and the arrowheadsindicate virus capsid proteins. (B) Heat-treated total extracts were subjected to Western blotting and probed with a polyclonal 4E-BP1 antibody.(C) Extracts prepared and blotted as in B were probed with 32P-labeled eIF4E. The analyses were performed as described.

the faster migrating form of 4E-BP1 was detected (Fig. 1C,lanes 11 and 12), further supporting the notion that thisinfection-induced 4E-BP1 species is the hypophosphorylatedform. Taken together, these results show that the kinetics ofdephosphorylation of 4E-BP1 temporally coincides with thatof the shutoff of the host protein synthesis and with theincreased binding of 4E-BP1 to eIF4E. Similar results werealso obtained with the EMCV R strain but with slower kineticsof dephosphorylation of 4E-BP1 (data not shown), in accor-dance with the finding that this virus causes a slower shutoff.

Translation Extracts Prepared from EMCV-Infected KrebsII Cells Do Not Reproduce the in Vivo Inhibition of Cap-Dependent Translation and 4E-BP1 Dephosphorylation. Be-cause eIF4E is necessary for cap-dependent but not forcap-independent translation, its sequestration by 4E-BP1would be expected to specifically inhibit cap-dependent trans-lation in an in vitro translation extract prepared from EMCV-infected cells. However, numerous earlier studies reported thatextracts prepared from EMCV-infected cells did not exhibitany selectivity in cap-independent versus cap-dependent trans-lation (15, 22, 25, 26). These findings are in apparent contra-diction with the in vivo data that point to a role for 4E-BP1 inthe mediation of the shutoff of host protein synthesis exertedby EMCV. To address this apparent discrepancy, we analyzed

E

A globin CAT EMC-CAT

'1 kDa(

w___1 3 4 5 6 789...... .

the status of 4E-BP1 phosphorylation in translation extractsprepared from EMCV-infected cells. Krebs II ascites transla-tion extracts were prepared either from mock-infected or 5-hpost-infected cells. Globin RNA and chloramphenicol acetyl-transferase (CAT) RNA (which are capped transcripts) andthe encephalomyocarditis (EMC)-CAT (which is uncappedand translates in a cap-independent fashion, driven by theEMCV internal ribosome entry site) have been used in thisstudy. In accordance with previous studies (22, 25, 26), allmRNAs tested, regardless of their dependency on the cap fortranslation, were translated less efficiently in extracts fromEMCV-infected than in mock-infected cells. The cap-dependent globin mRNA translation decreased 2.5- and 8-foldin EMCV R- and EMCV K-2-infected Krebs II ascites trans-lation extracts, respectively, as compared with translation inextracts from mock-infected cells (Fig. 2A, lanes 1-3). Trans-lation of CAT mRNA was also reduced 2.5- and 6-fold (lanes4-6; the position of the CAT protein is indicated by anarrowhead; the lower band is probably due to initiation at adownstream AUG). Significantly, cap-independent translationof EMC-CAT mRNA was also reduced to the same extent asglobin and CAT translation (lanes 7-9).The phosphorylation state of 4E-BP1 in the translation

extracts was monitored by Western analysis. No difference in

niRNA globin CAT EMC-CAT4E-BP1 - + + - +

EMC-CAT-in4EI Ui CAT

1 2 3 4 5 6

nmRNA globin CAT EMC-CAT4E-BP1 - + - + - +

EMC-CATCAT

1 2 3 4 5 6

FIG. 2. EMCV-infected cell extracts do not show selective cap-independent translation. (A) Extracts for translation were prepared from mock-or EMCV-infected (R and K-2 strain) Krebs II ascites cells and used to translate globin, CAT (cap-dependent), or EMC-CAT (cap-independent)mRNAs. Extracts were heat-treated and analyzed by Western blotting with a polyclonal anti-4E-BP1 antibody (inset). (B) Extracts prepared frommock-infected Krebs II ascites cells were used to translate globin, CAT, and EMC-CAT mRNAs in the presence or absence of 0.1 ,ug GST-4E-BP1.(C) Translation was performed as described for B, except that extracts were prepared from EMCV K-2-infected cells.

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

Proc. Natl. Acad. Sci. USA 93 (1996) 5581

the migration pattern of the different forms of 4E-BP1 be-tween mock-infected and 5-h post-infected cells was evidentwith the two different strains of EMCV (Fig. 2A, inset).Cap-dependent translation thus occurred in translation ex-tracts prepared from EMCV-infected cells, probably becauseof the absence of the dephosphorylated form of the repressor4E-BP1. The difference between the phosphorylation state of4E-BP1 in cell lysates and in in vitro translation extracts cantherefore explain the discrepancy between the in vivo and invitro patterns of protein synthesis after EMCV infection. Akinase activity that rephosphorylates 4E-BP1 could thereforebe active in translation extracts.An alternative, but less likely, explanation is that cap-

dependent translation in extracts prepared from ascites cells isresistant to inhibition by 4E-BP1. To rule out this possibility,the effects on translation of the addition of GST-4E-BP1 toascites extracts was examined. Addition of GST-4E-BP1 re-duced by 2.5-fold the translation of globinmRNA and by 2-foldthat of CAT mRNA (Fig. 2B, compare lanes 2 and 4 to lanes1 and 3). GST-4E-BP1 had only a marginal effect on thetranslation of EMC-CAT mRNA (Fig. 2B, lanes 5 and 6).Similar results were also obtained using a bicistronic RNA, inwhich the expression of the CAT is cap-dependent, and theexpression of the luciferase is cap-independent (data notshown). Furthermore, addition of GST-4E-BP1 to a transla-tion extract prepared from EMCV-infected ascites cells alsoselectively inhibited translation of cap-dependent mRNAs,although the extent of inhibition was smaller. The translationof globin mRNA and that of the CAT mRNA was decreasedby -1.5-fold (Fig. 2C, compare lanes 2 and 4 to lanes 1 and 3).In contrast, the translation of the EMC-CAT mRNA is notaffected by the addition of 4E-BP1 (Fig. 2C, compare lane 6 withlane 5). Taken together, these results indicate that 4E-BP1 isinhibitory in in vitro translation extracts prepared from ascitescells, provided it is dephosphorylated. Thus, the lack of prefer-ential inhibition of cap-dependent translation in vitro is probablyexplained by the absence of dephosphorylated 4E-BP1.4E-BP1 Is Underphosphorylated upon Infection of HeLa

Cells with Poliovirus. In poliovirus-infected cells, eIF4G israpidly cleaved (12, 41). This event precedes the shutoff of hostprotein synthesis. However, this by itself is not sufficient for

Alime (h) 2.0 2.5 3.0 3.5 4.0 4.5 5.0poliovirus - + - + - + - + - + - + - +

complete inhibition of protein synthesis (20, 21). We weretherefore interested to determine the fate of 4E-BP1 phos-phorylation in these cells. Cellular protein synthesis in polio-virus-infected HeLa cells was dramatically reduced as early as2.5 h after infection (Fig. 3A). The cleavage of eIF4G asassessed by Western blotting was first detected at 1.5 h afterinfection and was complete at 2.5 h (Fig. 3B, compare lanes 2and 6 to lanes 1 and 5), confirming that the cleavage of eIF4Gprecedes the shutoff of host protein synthesis. To analyze thestatus of 4E-BP1 phosphorylation following poliovirus infec-tion, a Western blot analysis of 4E-BP1 was performed. Inmock-infected cells (Fig. 3C, lane 1), three forms of 4E-BP1were detected by the antibody (indicated by arrowheads; notethat in human cells, three forms, as opposed to two species inmouse cells, are observed). At 4.5 h after infection (4 h in otherexperiments), the slower migrating band completely disap-peared, and the faster migrating bands became enhanced andshifted downwards on the gel (Fig. 3C, lanes 4 and 6). Todetermine whether the binding of 4E-BP1 to eIF4E wasaffected upon poliovirus infection, a Far-Western analysis wasperformed. In mock-infected HeLa cells, the eIF4E probe didnot interact significantly with 4E-BP1 (Fig. 3D, lane 1).Binding of the eIF4E probe to 4E-BP1 was strongly enhanced6 h after infection (Fig. 3D, lane 2), and the binding occurredwith the bands that were shifted downwards on the Westernblot shown in Fig. 3C. To further substantiate the conclusionthat the two novel bands, which are shifted downwards, are dueto 4E-BP1 dephosphorylation, 4E-BP1 was immunoprecipi-tated from 32PI-labeled cells. In mock-infected cells, threebands (Fig. 3E, lane 1, indicated by arrows) were detected. At3 h after infection, these three bands were still present.However, the slower migrating band disappeared 5 h afterinfection, and there was a global decrease of -70% in 32pincorporation (Fig. 3E, compare lane 3 with lane 1). Thus,poliovirus infection causes the dephosphorylation of 4E-BP1.However, the dephosphorylation occurs almost 2 h after theshutoff of host protein synthesis.Guanidine Prevents 4E-BP1 Dephosphorylation Induced by

Poliovirus Infection. Guanidine hydrochloride inhibits polio-virus replication (24, 42). In its presence, infection results in apartial shutoff of host mRNA translation (20). Strikingly,

Ctime (h) 3.0 4.5 6.0poliovirus - + - + - +

2kIDa X 4 fi

1 2 3 4 5 6

Dpohlovirus - +

- 2lkDi;

s.

2 3 4 5 6 7 8 910 1112 13 14

BInime (h) 1.5 2.0 2.5

poliovirus + + - +

1 2 3 4 5 6

2

Etime(h) 5 3 5poliovirus - + +

...

X 2 3

FIG. 3. Modification of 4E-BP1 following infection by poliovirus. (A) Time course of the shutoff. HeLa cells infected with poliovirus andmock-infected cells were labeled for 30 min with [35S]methionine at the indicated times after infection. Aliquots were analyzed by SDS/PAGE andautoradiography. (B) SDS/8% PAGE followed by Western blotting with an antibody against eIF4G was conducted using 25 ,ug of total protein.(C) Western analysis for 4E-BP1 was performed as described. (D) Far-Western analysis on 25 ,tg of protein, with a 32P-labeled eIF4E probe, wasconducted as described. (E) 32P-labeling of poliovirus-infected cells followed by immunoprecipitation with a polyclonal 4E-BP1 antibody.

Biochemistry: Gingras et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

5582 Biochemistry: Gingras et al.

guanidine hydrochloride does not prevent cleavage of eIF4G,suggesting that a second additional event is required for thecomplete shutoff of host protein synthesis (20). At 5.5 h afterinfection, in the presence of 1.5 mM guanidine hydrochloride,-60% of host cell protein synthesis was inhibited (Fig. 4A,compare lanes 2 and 4). Under these conditions, all the eIF4Gwas cleaved (Fig. 4B). The kinetics of eIF4G cleavage wassimilar in guanidine-treated cells as compared with untreatedcells (data not shown). However, the dephosphorylation of4E-BP1 did not occur in the presence of guanidine (Fig. 4C,compare lanes 2 and 4), suggesting that replication of polio-virus and subsequent translation are required for dephosphor-ylation of 4E-BP1.

DISCUSSIONThe best characterized mechanism for the shutoff of hostprotein synthesis by viruses is that used by poliovirus. Numer-ous studies showed that the poliovirus 2A protease is respon-sible for the inactivation of the cap-binding complex eIF4F bycleavage of its eIF4G component (for a recent review, see ref.12). However, cleavage of eIF4G does not occur followingEMCV infection. Here, we demonstrated that 4E-BP1, asuppressor of cap-dependent initiation of translation, is acti-vated by dephosphorylation following infection by EMCV,thus providing an attractive explanation for the shutoff of hostprotein synthesis by EMCV.The only initiation factor whose activity has been reported

so far to be modified by EMCV is eIF2a. This protein isphosphorylated by the double-stranded RNA-dependent ki-nase PKR as a result of infection with several viruses (43).Phosphorylation of eIF2a would be expected to cause ageneral inhibition of translation rather than a specific shutoffof host protein synthesis, because this factor is required for theformation of the Met-tRNA ternary complex and ribosomebinding to all mRNAs, including viral RNAs. However, it hasbeen postulated that phosphorylation of eIF2a results in astronger competition between cellular and viral RNA for thereduced amount of eIF2-GTP-Met-tRNA complex, and thiscompetition was suggested to explain in part the shutoff (16,44, 45). Another hypothesis posits that the changes in theintracellular ion concentrations upon EMCV infection selec-tively inhibit cellular mRNA translation (46). Our resultsprovide a simple, alternative, and attractive explanation for thepreferential translation of EMCV RNA in infected cells.

Apoliovirus - + - +guanidine - - + +

NM .;i .";

*-

:|

A difficulty that has seriously hampered the studies on themechanism of the shutoff of host protein synthesis by EMCVis that extracts prepared for in vitro translation do not reca-pitulate the in vivo discrimination against capped cellularmRNAs (15, 16, 46). In this study, we did not observe aselective translation of viral RNA either. However, the in vitrotranslation extracts from EMCV-infected cells did not containthe underphosphorylated form of 4E-BP1. This agrees wellwith our model that dephosphorylation of 4E-BP1 is respon-sible for the shutoff, as the pattern of 4E-BP1 forms intranslation extracts prepared from infected cells is indistin-guishable from that prepared from mock-infected cells. It isnoteworthy that addition of bacterially expressed 4E-BP1(unphosphorylated) to Krebs extracts, infected or not infectedwith EMCV, was sufficient to selectively inhibit cap-dependenttranslation, as was first observed in reticulocyte lysate. There-fore, the lack of effect seen in the in vitro extracts is really dueto the absence of the underphosphorylated 4E-BP1 in thissituation. Similarly, extracts prepared from poliovirus-infectedcells did not contain underphosphorylated 4E-BP1 (A.-C.G.,unpublished results), although they translate poorly cappedcellular mRNAs. This translational inhibition is due to cleav-age of the eIF4G. Thus, it appears that the dephosphorylationof4E-BP1 in infected cells is a reversible phenomenon and thatthe 4E-BP1 becomes readily rephosphorylated upon prepara-tion of in vitro translation extracts. Indeed, translation extractsare traditionally prepared in buffers lacking any kinase orphosphatase inhibitors, as opposed to the extracts used for theWestern and Far-Western assays. Some of the inhibitorspresent in extracts prepared for Western analysis (NaF, sodiumorthovanadate) are known to inhibit translation and cannot beused in the preparation of in vitro translation extracts (47).4E-BP1 dephosphorylation following poliovirus infection

lags behind the shutoff of host protein synthesis and thereforecannot explain it. It is possible, however, that in poliovirus-infected cells, the dephosphorylation of 4E-BP1 accentuatesthe inhibition of host mRNA translation late in infection.Alternatively, it is also conceivable that the dephosphorylationof 4E-BP1 does not play a role in the shutoff of host proteinsynthesis by poliovirus but is a relic of evolution. Dephosphor-ylation of 4E-BP1 might have been used by all picornavirusesearly in their evolution, but cleavage of eIF4G has evolved forsome picornaviridae family members as a more efficient meansto inhibit host protein synthesis. The two events, the cleavageof eIF4G and the dephosphorylation of 4E-BP1, share acommon target, the cap recognition process.

Bpoliovirusguanidine

- +§ - +

_- + +

l 2 3 4

Cpoliovirusguanidine

2lkDa -

1 2 3 4

+ +

_- + +

1

l 2 3 4

FIG. 4. Effect of guanidine hydrochloride on eIF4G cleavage and 4E-BP1 dephosphorylation following poliovirus infection. HeLa cells wereinfected with poliovirus for 5 h in the presence or absence of guanidine hydrochloride (1.5 mM). Guanidine hydrochloride was added immediatelyfollowing virus adsorption. (A) Shutoff of protein synthesis was monitored by [35S]methionine labeling as described. (B) Extracts from cells treatedas in A were analyzed by SDS/6% PAGE followed by Western blotting with an antibody against eIF4G. (C) Cell extracts as in A and B wereheat-treated and analyzed by SDS/15% PAGE followed by Western blotting for 4E-BP1 as described.

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0

Proc. Natl. Acad. Sci. USA 93 (1996) 5583

The dephosphorylation of 4E-BP1 is a late event in polio-virus infection that requires RNA replication and extensivevirus protein expression, because it is inhibited by guanidine.This differs from eIF4G cleavage, which occurs in the absenceofvirus replication. It is not established that the lack of4E-BP1dephosphorylation in the presence of guanidine can accountfor the residual cellular protein synthesis, because 4E-BP1dephosphorylation occurs after the shutoff of host proteinsynthesis in the absence of guanidine. The dephosphorylationof 4E-BP1 is not likely to be the result of the inhibition ofprotein synthesis, because inhibition of host protein synthesisby adenovirus or vaccinia virus did not lead to dephosphory-lation of 4E-BP1 (A.-C.G., unpublished observations).The involvement of 4E-BP1 in EMCV-mediated shutoff of

host protein synthesis reinforces the idea that even relatedviruses such as EMCV and poliovirus have evolved differentmechanisms to maximize their translation and that the shutoffis a complex event involving many factors. However, our datado not exclude the contribution of other mechanisms, such ascompetition between RNAs or cellular permeabilization anddifferences in ionic conditions that favor the translation ofviral mRNAs (47). The mechanism of dephosphorylation of4E-BP1 in infected cells is currently under investigation todetermine whether a specific phosphatase is activated orwhether a kinase pathway is impaired after virus infection.

We thank members of the laboratory for helpful comments on themanuscript. We thank Dr. A. Palmenberg for the EMCV R strain andDr. L. Carrasco and Isabel Novoa for the a-eIF4G antibodies. Thiswork was supported by grants from the National Cancer Institute ofCanada and the Medical Research Council of Canada to N.S. andNorth Atlantic Treaty Organization Grant CRG 950222 to G.J.B andN.S.; A.-C.G. was supported by an National Science and EngineeringResearch Council 67 Studentship, and A.P. was supported by a CancerResearch Society of Montreal Studentship.

1. Mathews, M. B., Sonenberg, N. & Hershey, J. W. B. (1996) inTranslational Control, eds. Hershey, J. W. B., Mathews, M. B. &Sonenberg, N. (Cold Spring Harbor Lab. Press, Plainview, NY),pp. 1-29.

2. Shatkin, A. J. (1976) Cell 9, 645-653.3. Duncan, R., Milburn, S. C. & Hershey, J. W. (1987)J. Biol. Chem.

262, 380-388.4. Hiremath, L. S., Webb, N. R. & Rhoads, R. E. (1985) J. Biol.

Chem. 260, 7843-7849.5. Rozen, F., Edery, I., Meerovitch, K., Dever, T. E., Merrick, W. C.

& Sonenberg, N. (1990) Mol. Cell. Biol. 10, 1134-1144.6. Edery, I., Humbelin, M., Darveau, A., Lee, K A., Milburn, S.,

Hershey, J. W., Trachsel, H. & Sonenberg, N. (1983) J. Biol.Chem. 258, 11398-11403.

7. Ray, B. K., Lawson, T. G., Kramer, J. C., Cladaras, M. H., Grifo,J. A., Abramson, R. D., Merrick, W. C. & Thach, R. E. (1985) J.Biol. Chem. 260, 7651-7658.

8. Pelletier, J. & Sonenberg, N. (1988) Nature (London) 334,320-325.

9. Jang, S. K., Krausslich, H. G., Nicklin, M. J., Duke, G. M., Pal-menberg, A. C. & Wimmer, E. (1988) J. Virol. 62, 2636-2643.

10. Jackson, R. J., Howell, M. T. & Kaminski, A. (1990) TrendsBiochem. Sci. 15, 477-483.

11. Sonenberg, N. (1987) Adv. Virus Res. 33, 175-204.

12. Ehrenfeld, E. (1996) in Translational Control, eds. Hershey, J. W.B., Mathews, M. B. & Sonenberg, N. (Cold Spring Harbor Lab.Press, Plainview, NY), pp. 549-573.

13. Leibowitz, R. & Penman, S. (1971) J. Virol. 8, 661-668.14. Kaufmann, Y., Goldstein, E. & Penman, S. (1976) Proc. Natl.

Acad, Sci. USA 73, 1834-1838.15. Abreu, S. L. & Lucas-Lenard, J. (1976) J. Virol. 18, 182-194.16. Svitkin, Y. V., Ginevskaya, V. A., Ugarova, T. Y. & Agol, V. I.

(1978) Virology 87, 199-203.17. Helentjaris, T. & Ehrenfeld, E. (1978) J. Virol. 26, 510-521.18. Scheper, G. C., Voorma, H. 0. & Thomas, A. A. (1992) J. Biol.

Chem. 267, 7269-7274.19. Pause, A., Methot, N., Svitkin, Y., Merrick, W. C. & Sonenberg,

N. (1994) EMBO J. 13, 1205-1215.20. Bonneau, A.-M. & Sonenberg, N. (1987) J. Virol. 61, 986-991.21. Perez, L. & Carrasco, L. (1992) Virology 189, 178-186.22. Jen, G. & Thach, R. E. (1982) J. Virol. 43, 250-261.23. Mosenkis, J., Daniels-McQueen, S., Janovec, S., Duncan, R.,

Hershey, J. W. B., Grifo, J. A., Merrick, W. C. & Thach, R. E.(1985) J. Virol. 54, 643-645.

24. Rose, J. K., Trachsel, H., Leong, K. & Baltimore, D. (1978) Proc.Natl. Acad. Sci. USA 75, 2732-2736.

25. Lawrence, C. & Thach, R. E. (1974) J. Virol. 14, 598-610.26. Svitkin, Y. V., Ugarova, T. Y., Ginevskaya, V. A., Kalinina,

N. O., Scarlat, I. V. & Agol, V. I. (1974) Intervirology 4, 214-220.27. Pause, A., Belsham, G. J., Gingras, A.-C., Donze, O., Lin, T. A.,

Lawrence, J., Jr., & Sonenberg, N. (1994) Nature (London) 371,762-767.

28. Lin, T. A., Kong, X., Haystead, T. A., Pause, A., Belsham, G.,Sonenberg, N. & Lawrence, J., Jr. (1994) Science 266, 653-656.

29. Hu, C., Pang, S., Kong, X., Velleca, M. & Lawrence, J., Jr. (1994)Proc. Natl. Acad. Sci. USA 91, 3730-3734.

30. Mader, S., Lee, H., Pause, A. & Sonenberg, N. (1995) Mol. Cell.Biol. 15, 4990-4997.

31. Haghighat, A., Mader, S., Pause, A. & Sonenberg, N. (1995)EMBO J. 14, 5701-5709.

32. Rueckert, R. R. & Pallansch, M. A. (1981) Methods Enzymol. 78,315-325.

33. Dalgarno, L., Martin, E. M., Liu, S. L. & Work, T. S. (1966) J.Mol. Biol. 15, 77-91.

34. Laemmli, U. K. (1970) Nature (London) 227, 680-685.35. Blanar, M. A. & Rutter, W. J. (1992) Science 256, 1014-1018.36. Svitkin, Y. V. & Agol, V. I. (1978) FEBS Lett. 87, 7-11.37. Svitkin, Y. V., Lyapustin, V. N., Lashkevich, V. A. & Agol, V. I.

(1984) Virology 135, 536-541.38. Graves, L. M., Bornfeldt, K. E., Argast, G. M., Krebs, E. G.,

Kong, X. M., Lin, T. A. & Lawrence, J. C. (1995) Proc. Natl.Acad. Sci. USA 92, 7222-7226.

39. Lin, T. A., Kong, X. M., Saltiel, A. R., Blackshear, P. J. &Lawrence, J. C. (1995) J. Biol. Chem. 270, 18531-18538.

40. Beretta, L., Gingras, A.-C., Svitkin, Y. V., Hall, M. N. & Sonen-berg, N. (1996) EMBO J. 15, 658-664.

41. Etchison, D., Milburn, S. C., Edery, I., Sonenberg, N. & Hershey,J. W. (1982) J. Biol. Chem. 257, 14806-14810.

42. Bablanian, R. (1972) Virology 47, 255-259.43. Black, T. L., Safer, B., Hovanessian, A. & Katze, M. G. (1989) J.

Virol. 63, 2244-2251.44. Baglioni, C., Simili, M. & Shafritz, D. A. (1978) Nature (London)

275, 240-243.45. Duke, G. M., Hoffman, M. A. & Palmenberg, A. C. (1992) J.

Virol. 66, 1602-1609.46. Carrasco, L. & Smith, A. E. (1976) Nature (London) 264, 807-

809.47. Carrasco, L. (1994) Pharmacol. Ther. 64, 215-290.

Biochemistry: Gingras et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

0