Isolation and structural characterization of cap-binding proteins from poliovirus-infected HeLa cells.

0022-538X/85/050515-10$02.00/0

Copyright C 1985,American Society forMicrobiology

Isolation and Structural

Characterization of Cap-Binding Proteins

from Poliovirus-Infected HeLa Cells

KEVIN A. W. LEE, ISAAC EDERY, AND NAHUM SONENBERG*

Department of Biochemistry and McGill Cancer Center, McGill University, Montreal, Quebec, Canada H3G 1Y6 Received 2 August 1984/Accepted 8 February 1985

Inpoliovirus-infected HeLa cells, poliovirus RNA is translatedattimes whencellularmRNA translationis strongly inhibited. It is thought that this translational control mechanism is mediated by inactivation of a cap-binding protein complex (comprising polypeptidesof 24[24-kilodaltoncap-binding protein], 50,and -220 kilodaltons). Thiscomplexcanrestorethe translation ofcapped mRNAs in extracts frompoliovirus-infected cells. We havepreviouslyshown that thevirally induced defect prevents interaction betweencaprecognition factorsand mRNA.Here,weshowthatthecap-binding protein complex (andnotthe 24-kilodaltoncap-binding protein)hasactivitythatrestoresthecap-specific mRNA-proteininteraction when added to initiation factors frompoliovirus-infected cells. Thus, theactivitythatrestores thecap-specificmRNA-proteininteractionand that whichrestoresthe translation of cappedmRNAs inextractsfrom poliovirus-infected cells, copurify.The results also indicate, byanalternativeassay,that the cap-bindingprotein complex istheonlyfactor inactivated by poliovirus. We also purified cap-binding proteins from uninfected andpoliovirus-infected HeLa cells. By various criteria, the 24-kilodalton cap-binding protein is not structurally modified as a result of infection. However,the220-kilodaltonpolypeptideof thecap-binding protein complexisapparentlycleavedbyaputative viral (orinduced) protease. Byin vivo labelingandm7GDPaffinity chromatography, we isolated amodified cap-binding protein complex frompoliovirus-infected cells, containing proteolytic cleavage fragments of the 220-kilodalton polypeptide.

The cap structurem7GpppX(m)isfoundatthe 5' terminus

of almost all eucaryotic mRNAs, picornaviral and some plant viralRNAs being notable exceptions(27). Many stud-ies have indicated that the cap structure facilitates 40S

ribosome attachment to mRNA during initiation of transla-tion (2, 27), and it was anticipated that this function is

mediated by acap-specific mRNA-protein interaction.

By chemical cross-linking to 3H-labeled oxidized capped viral mRNAs, it has been shown that polypeptides of 24, 50, and 80 kilodaltons (kDa) present in crude initiation factors (IF) fromrabbitreticulocytes and severalother mammalian sources (12, 19, 28; K. A. W. Lee and N. Sonenberg, unpublished data) specifically interact with the cap struc-ture. The

identity

oftwoofthese

polypeptide

isknown: the

24-kDa polypeptide corresponds to the

24,000-molecular-weight(24K)cap-binding protein (CBP) (28, 29; see below), and the50-kDapolypeptide correspondstoeIF-4Abasedon thefactthat itcanbeimmunoprecipitatedwith a monoclonal

antibody againsteIF-4A(6) and that purified eIF-4A canbe

cross-linked to mRNA with characteristics similar to those of the 50-kDapolypeptide (6, 10). Theidentityof the 80-kDa

polypeptide is not established, but several results strongly suggest that it is eIF-4B. It was shown that this factor can

specifically cross-linkto5'oxidized reovirus mRNA only in the presence of

ATP-Mg2+

(6, 10) as demonstrated for the 80-kDapolypeptideinpreparations of crude IF. In addition,

cross-linking of eIF-4B requiresthe presence ofother initi-ation factors (eIF-4A [10] and CBP complex [6]).

Polypeptideswith affinityfor the cap structure have been

purified from rabbit reticulocytes by

m7GDP

affinity chro-matography. Originally, a 24K cap-binding protein

(24K-CBP, CBPI, or eIF-4E) was purified by Sonenberg et al.

(30), and subsequently a high-molecular-weight complex

comprisingthe 24K-CBP andmajor polypeptides of 50 and

* Correspondingauthor.

220 kDawaspurifiedby several groupsindependently(6,11, 31). This complex is referred to as CBP II, eIF-4F, orthe

CBPcomplex (throughoutthispaper). The 50-kDa

polypep-tide is very similarto eIF-4A asdetermined by

two-dimen-sional gel analysis (11), peptide map analysis (6), and im-munoreactivity (6). Furthermore, both polypeptidesexhibit

similar mRNA cross-linking characteristics, and we will therefore refer to this polypeptide as eIF-4A. Experiments

with purified factors have shown that the CBP complex,

eIF-4B, and mRNAaresufficienttoreconstitutethe

cap-spe-cific mRNA-protein interaction observed by the chemical

cross-linking assay when crude initiation factors are used (6). These results suggest thatinteraction of the CBP com-plex and eIF-4B with the capstructure somehowfacilitates attachmentof 40S ribosomal subunits to capped mRNAs.

Poliovirus infection ofHeLa cells results in a rapid and

apparently quantitative inhibition of cellularprotein synthe-sis, such that viral RNA is almost exclusively selectedfor translation (1, 8). It was shown that crude IF from

polio-virus-infected cells could stimulate translation ofpoliovirus RNAin vitro buthadnosucheffectontranslation of cellular mRNAs(15). Consequently, various groups wereledto ask which particular initiation factorwas inactivated by polio-virus.Using different approaches, Helentjaris etal. (16)and Rose et al. (25) obtained evidence that eIF-3 and eIF-4B, respectively, were inactivated.

Theseapparentlyconflicting observationswere soon tobe reconciled. On discovery that poliovirus RNA is naturally uncapped (17, 22), it was an attractive hypothesis that inactivation of some form of CBP actually explains the shut-offofhost protein synthesis (33). In accord with this

idea, Tahara et al. (31) purified a protein complex (-8 to 10S) by m7GDP affinity chromatography which comprised

major polypeptides of24 (24K-CBP) 50, 94, and -220 kDa and which could restore translation ofcapped mRNAs in extractsfrompoliovirus-infected cells. This latteractivity is 515

on November 10, 2019 by guest

http://jvi.asm.org/

referred to as restoring activity. Recently, we have de-scribed a CBP complex (the CBP complex [6]) comprising

the 24K-CBP and polypeptides of 50 (eIF-4A) and -220

kDa,which also has restoringactivity(7). The24K-CBPcan

also be detected in preparations ofeIF-3andeIF-4B (29), as canother CBPcomplex components (9, 11), thus mostlikely explaining theeffects previously attributed to these factors

(16, 25). By a different approach (19), we analyzed CBPs

after poliovirus infection, using the chemical cross-linking assay, and showed that thecap-binding activity ofthe 24-,

50-, and80-kDacap-specific polypeptides was almost

com-pletely abolished after infection. In contrast tothis, Hansen and Ehrenfeld (12), by using the cross-linking assay,

re-ported no change in theamount ofthe 24-kDa polypeptide but did find that the 24-kDa polypeptide no longer

cosedi-ments with eIF-3 after poliovirus infection (13). These results suggested a modification to CBP which possibly prevents afunctional association between eIF-3and CBP.

Taken together, these results engender the belief that someform ofCBPis indeed inactivatedbypoliovirus,and a reportfrom Etchison et al. (9) pointed to the likely

mecha-nism. Using antisera against a 220-kDa polypeptide (P220) present inpreparations ofeIF-3 (underconditions in which

therestoring activity fractionates with eIF-3), these authors

showed that P220 is degraded in poliovirus-infected cells. Theanti-P220antibody also recognizes the -220-kDa

poly-peptide oftheCBPcomplex(9),andsoitwasproposedthat

proteolysis ofP220 by a poliovirus-dependent protease re-sults inthe shut-off ofhost protein synthesis.

Here we show that addition ofthe CBP complex to IF

from poliovirus-infected HeLa cells (I-IF) can restore the

interaction betweenthe 80-kDapolypeptide(present inI-IF) and the cap structure, asassayed bythechemical cross-link-ing technique. The 24K-CBP has no suchactivity, strongly suggestingthatrestoration ofthe80-kDacap-bindingactivity

and restoration ofcapped mRNAfunction in extractsfrom poliovirus-infected cellsare due to the same activity. In an attempt to demonstrate directly the defect in the CBP complex caused by

poliovirus,

we isolated CBPs from uninfected and poliovirus-infected HeLa cells. By various criteria, the 24K-CBP is unaltered by

poliovirus infection,

whereas the CBP complex is structurally modified. By in

vivo labelingof cells andsubsequent

m7GDP

affinity purifi-cation ofCBPs, we obtained aCBP complex from

poliovi-rus-infectedHeLacellswhich containsproteolytic cleavage fragments ofP220.

MATERIALSAND METHODS

Cells and virus. Mouse L-929cells and HeLaS3 cellswere grown insuspension in 10% calfserum. Infection ofLcells with reovirus type 3 (Dearing strain; 10 PFU per cell) and

virus purification were performed as previously described

(3). Infection of HeLa cells withpoliovirustype 1(Mahoney strain; 10 to 20 PFU percell, except where otherwise

indi-cated) was bythe method of Rose et al. (25).

Preparation of

methyl-3H-labeled

oxidized reovirusmRNA. Synthesisof

methyl-3H-labeled

reovirus mRNA to aspecific activityof -8 x 104cpm/,ugwithviral cores in thepresence

ofS-adenosylmethionine

(specific activity,

70

Ci/mmol;

New EnglandNuclearCorp.;1Ci = 3.7x

1010

Bq) andperiodate

oxidation were bythe method of Muthukrishnan et al. (21).

Preparation of crude protein synthesis initiation factors.

Preparation of rabbit reticulocyte

lysate,

high-salt wash of ribosomes (as a source of initiation factors), and subfraction-ation to a 0 to 40% ammonium sulfate fraction were as described by Schreier and Staehelin (26). Preparation of

HeLa cell extracts and crude initiation factors was by the

method of Lee and Sonenberg (19).

Cross-linkingof mRNAtoprotein synthesisfactors.

methyl-3H-labeled oxidized reovirus mRNA was incubated with

crude IForpurifiedCBPsorboth(asdescribed in thefigure legends) for 10 minat 30°C essentiallyas described before

(19), followed by addition of NaBH3CN

(freshly prepared

solution; AldrichChemical

Co.,

Inc.)

overnight

and RNase A to digest the mRNA. The samples were resolved on

sodium dodecyl sulfate

(SDS)-polyacrylamide

gels, and la-beled bandsweredetected

by fluorography.

All

cross-linking

incubationswerecarriedoutin thepresenceof

ATP-Mg2'

as previously described (19).

Purification of CBPs. Purification of rabbit

reticulocyte

CBP

complex

was

essentially

as described

by Edery

et al.

(6). A 0 to 40% ammonium sulfate fraction of ribosomal

high-salt wash was layered on 12-ml, 10 to 35% linear

sucrose

gradients

in buffer A (20 mM HEPES

[N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid] [pH 7.5],

0.2 mM

EDTA,

0.5 mM

phenylmethylsulfonyl fluoride,

and 7 mM

3-mercaptoethanol) containing

0.5 M KCI. Centrifu-gation wasfor24 h at 38,000rpm in an SW40rotor at

4°C.

The tophalf ofthe

gradient, excluding

the

fast-sedimenting

eIF-3

(>1OS),

was

pooled

and

dialyzed against

buffer A

containing

0.1 M KCI and 10%

glycerol.

The

dialyzed

material was then loaded

directly

onto an

m7GDP-agarose

affinity

column(6,

9)

equilibrated

inbufferA

containing

0.1 M KCI and 10%

glycerol.

Nonspecifically

bound

proteins

were eluted

by washing

the column in 50 ml of

buffer

A

containing

0.1M KCIand10%

glycerol,

followed

by

4mlof

100 ,IMGTP in thesamebuffer.

Cap-specific proteins

were thenelutedwith75 ,uM m7GTPinbufferA

containing

0.1 M

KCIand 10%

glycerol.

Rabbit

reticulocyte

24K-CBP was

purified

fromthe S100 fraction

by

a modification ofthe

procedure

ofTaharaetal.

(31). S100 was mixed with DEAE-cellulose (3 volumes of S100and 1volumeof swollen

DEAE-cellulose)

equilibrated

in low-column buffer

(LCB;

20 mMTris

[pH 7.5],

0.2 mM

EDTA,

7 mM

P-mercaptoethanol)

containing

80 mM KCI.

The 24K-CBP binds to

DEAE-cellulose

underthese condi-tions. The resin was then washed

extensively

with LCB

containing

80 mM KCl to remove excess

hemoglobin,

fol-lowed

by

batchelution ofbound material with LCB

contain-ing

250 mM KCl.The eluate was thenconcentrated

by

0to 50% ammonium sulfate

fractionation,

the

precipitate

was

dialyzed

against

LCB

containing

200 mM

KCl,

and the

dialyzed

material was diluted twofold before

being

loaded ontothe

m7GDP

affinity

column.

Purification of CBP

complex

from HeLa cells was very similarto the

purification

ofrabbit CBP

complex,

with the

exception

that the crude IFs were not fractionated

by

ammonium sulfate. For each

purification,

not less than 20 litersof

log-phase

HeLacellsat acell

density

of5x 105cells per mlwasusedtoprepare crude IF. In the caseofinfected

lysates,

infection was verified

by

various

criteria,

e.g., the presence of viral

antigens by

immunoblotting

and the trans-lational

specificity (capped

versus

naturally

uncapped

mRNA

translation)

of cell extractsin in vitro translation.

Purification of24K-CBPfrom HeLacellswasachieved

by

passing

total

post-ribosomal

supernatant

(S100)

over the

m7GDP affinity column,

followed

by

elution as described

above for the CBP

complex.

Purification of CBP complex from in vivo-labeled HeLa cells.HeLacells

(13

mlat4 x

105

cellsper

ml)

were

pelleted

and

suspended

in 8 mlof methionine-free medium

containing

20%

dialyzed

fetalcalfserumand 200

,uCi

of

[35S]methionine

on November 10, 2019 by guest

http://jvi.asm.org/

per ml (.1,000Ci/mmol, New England Nuclear). Labeling

was for6 h,after which thecells were split equally in two.

Halfwere infected with poliovirus inavolumeof400

[LI

at a multiplicity of infection of 50 PFU per cell, and the other half were mock infected. The conditions of adsorption and

infection were as described by Rose et al. (25). At 3 h

postinfection, cells werepelletedandsuspendedin 180

p1

of lysis buffer containing 150 mMNaCl, 20 mM Tris (pH 7.5), 0.5% Nonidet P-40, and 2 mM

phenylmethylsulfonyl

fluo-ride. The suspension was adjusted to 600 mM potassium

acetateandthenlefttostironice for 30min.CrudeIF(-200

,ug)from uninfectedcells wasthen addedascarrier,and the

cell extracts were sedimented through a linear 10 to 30%

sucrose gradient in buffer A containing 0.5 M potassium

acetate in an SW50.1 rotor at 39,000 rpmfor 15 h. Catalase (llS) was run on a separate gradient, and the material migrating slowerthancatalase inthegradientwaspooledfor

m7GDP affinity chromatography. Thepooled fractions were dilutedwithwater to afinalpotassiumacetateconcentration of100 mM, and this material was loaded

directly

onto the

affinity column. The columnwasthen washed with 50 ml of LCB containing 100 mM KCI andthen with 20 ml of100

F.M

GDP in LCBcontaining 100 mMKCl.Thefirst 1 mlof GDP eluate wascollected. The

affinity

resinwasthen transferred

to an Eppendorf tube, and the

cap-specific polypeptides

were batch eluted with 100

p.M m7GDP

in LCB containing

100 mM KCl.

Preparation of polyclonal antisera

against sheep

CBP complex was carried out

by

the method of Vaitukaitis

(34).

The CBP complex was purified from sheep erythrocytes under the protocol described for the

purification

of the

24K-CBP from rabbit reticulocyte S100 (see above). CBP complex (20

p.g)

in LCB

containing

500 mMKCI wasmixed

with 1.2 volumes ofcomplete Freund adjuvant. This

mate-rialwasinjectedintradermally intotheback ofarabbitinca. 20differentspots. Fourmonths

later,

therabbitwasboosted subcutaneously with 20

p.g

of

antigen injected

in three

different places inthe back. One week

later,

the rabbit was bled through the earand serum was

prepared.

Immunoblot analysis (Western

blotting)

was

performed

essentially as described by

Edery

et al. (6).

Polypeptides

were resolved on a10% SDS polyacrylamide

gel

and trans-ferredtonitrocellulose paper, and the blotwasincubatedin

Tris-buffered saline (pH 7.5)

containing

1% bovine serum albumin for 1 h. This was followedby incubation

overnight

inTris-buffered saline

containing

1% bovine serum albumin and the anti-CBP complex

antibody.

Blots were washed in

Tris-buffered saline andthenwereincubated with

peroxidase-conjugatedgoatanti-rabbit

immunoglobulin G;

immunoreac-tive species were visualized by color development with diaminobenzidene (32). Antisera

against

P220

(prepared

as described above) were diluted 1:330 in Tris-buffered saline

containing 1% bovine serumalbuminbefore incubation with nitrocellulose blots.

RESULTS

Taharaetal.(31)and,morerecently, Ederyetal.(7)have shown that the activity that restores translation ofcapped

mRNAs in extracts from

poliovirus-infected

HeLa cells

(referred to as restoring activity) copurifies with the CBP

complex but does not reside in the 24K-CBP (the 24-kDa subunitoftheCBPcomplex).Usingadifferentapproach,we analyzedCBPs in IF from poliovirus-infectedcells(I-IF) by

thechemicalcross-linkingassay and detected reduced levels ofthe 24-, 50-,and 80-kDapolypeptides compared with the levelsdetected in IF from uninfected cells (U-IF) (19).

1 2 3 4 5 6 7 8

i5.

111

80K.ab - - -- MO

A44

. AM -l _f _f AMd_

5OKt_--I _ ~~ a~_ft_ Vm_

24K

0

0

9 10 11 12

I

_ As

_m

_

4b

.'P-f +- + - + + -4 - 4

FIG. 1. Effects of purified CBPs on chemical

cross-linking

of crude IFfrompoliovirus-infectedcells.CrudeIF,purified CBPs,or mixtures ofthe two wereincubated undercross-linkingconditions with 3H-oxidizedreovirusmRNAasdescribed in thetext. Labeled polypeptideswere then resolvedon 10%

SDS-polyacrylamide gels

and autoradiographed. Lanes 1 and 2 contained -100 ,ug of crude U-IF;lanes3and4, -100,ug of crude U-IFplus1.5 ,ug

(containing

-0.3 ,ug of24K-CBP)of rabbitreticulocyteCBP

complex;

lanes 5 and 6, -100 ,ug ofcrudeI-IF;lanes7and8, -100 ,ugof crudeI-IF plus1.5 ,ug(containing-0.3 ,ug of24K-CBP)of rabbitreticulocyte CBPcomplex;lanes 9and10, 2,ugof rabbitreticulocyte24K-CBP fromthe S100fraction;lanes 11 and12, -100 ,ugof crudeI-IFand 2,g of rabbit reticulocyte 24K-CBP. m7GDP (0.67 mM) was includedorexcludedas indicated below thefigure.

We wanted to test the

hypothesis

that the activities

required

to restore

capped

mRNAfunction inextractsfrom

poliovirus-infected

cells and for the interaction of the

cap-specific polypeptides

with the cap structure

(as

assayed by

cross-linking)

are identical and reside in the CBP

complex.

Tothis

end,

we

assayed

the

ability

of

purified

CBP

complex

to restore

cap-specific cross-linking

ofthedifferent

cap-spe-cific

polypeptides.

We have

previously

demonstrated that

cross-linking

of the CBP

complex by

itself,

in the presence or absenceof

ATP-Mg2+,

results in

cap-specific

cross-link-ing

of the 24K-CBP

only

(6;

the same

preparation

of CBP

complex

was used in these

experiments).

Addition of

puri-fied eIF-4B to the CBP

complex (in

the presence of

ATP-Mg2+)

results in

cap-specific cross-linking

of eIF-4A and eIF-4Bin additiontothe24K-CBP

(6).

These results

strongly

suggest that the 80-kDa

polypeptide

thatcanbecross-linked

incrude IF is eIF-4B.

Figure

1

(lane

1)

shows the

cross-link-ing

profile

ofa total IF

preparation

from uninfected HeLa cells.

Cross-linking

of several

polypeptides

is inhibited

by

the addition of m7GDP (compare lane 2 with lane

1),

as

previously reported by

us

(19),

andincludesthe

24-, 50-,

and

80-kDa

polypeptides.

The 24-kDa

polypeptide

indicated in the

figure

is

actually

adoubletwhich

corresponds

to the 26-and28-kDa

cap-specific

polypeptides

described in reference 19. We also

previously

reported

specific

cross-linking

ofa

polypeptide

of -32 kDa in U-IF

(19),

whichcanalsobeseen in

Fig. 1,

lane 1. The amount of this

polypeptide

varies in different

preparations,

however,

and is often

completely

absent. Its

significance,

if any, is therefore notclear. Addi-tionof the CBP

complex

to U-IFhad no

stimulatory

effect on

cross-linking

of the 80-kDa

cap-specific

polypeptide

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.612.322.555.77.271.2]

present in the IF

preparation,

whereas there was a small increase in the amount of cross-linked 24- and 50-kDa

cap-specific polypeptides (twofold,

asdetermined

by

densi-tometryof the labeled bands;compare lanes 3 and 1 inFig.

1).

Cross-linking

ofI-IF resulted in a very small amount of

specific cross-linking

of the 24-kDa

polypeptide only,

as

previously reported

(19;

Fig.

1, lanes 5 and 6). However,

addition of theCBP

complex

toI-IFrestored the

cross-link-ing profile

to that observed when U-IF was used alone

(compare lanes 7 and 1 in

Fig.

1; the

24-, 50-,

and 80-kDa

cap-specific polypeptides

areindicated

by

arrowheadstothe

right

oflane

8).

Itisnot

possible

totell from the

cross-linking

in lane 7whether thecross-linked 24-and 50-kDa

polypep-tides arecontributed

by

theI-IFortheCBPcomplex, since both the 24- and 50-kDa

polypeptides

are present in both fractions.

However,

it is clear that the

activity required

for

the

cap-specific cross-linking

of the 80-kDa

polypeptide

is

presentinthe added CBP

complex

andis

lacking

intheI-IF

preparation.

Since the 80-kDa

polypeptide

is not presentin theCBP

complex

(6),the

only

interpretation

ofthis

experi-mentis that

(i)

the80-kDa

polypeptide (probably

eIF-4B) is notinactivated in

poliovirus-infected

cellsasassayed bythe

cross-linking

assay, in accord with earlier reports (5, 16),

and (ii) it cannot cross-link to mRNA in I-IF because the

CBP

complex

is inactivated.Ourresultsalsoindicate

(based

onanassayother than restoration of

capped

mRNAfunction

in extracts from

poliovirus-infected

cells) that the CBP

complex

is the

only

initiation factor inactivated

during

poliovirus

infection. It was also

important

to assay the

ability

of

purified

24K-CBPto restorethe

cap-specific

cross-linking

of the 80-kDa

polypeptide

in I-IF in

light

of data

showing

that it

might

be

required

for eIF-4B

cross-linking

(10).

The

cross-linking

ofthe24K-CBPis

completely

sensi-tive to

m7GDP

(compare lanes 9 and 10 in

Fig.

1).

Addition

of 24K-CBP toI-IFresulted in

specific cross-linking

ofthis

polypeptide,

albeit at a somewhat reduced level (compare lanes12 and 10),

presumably

dueto

competition

forlabeled

oxidized mRNA from the vast excess of

non-cap-specific

polypeptides

present in the I-IF. Addition ofthe

24K-CBP,

however,

did not enhance the

cross-linking

of any other

polypeptide

in the I-IF

preparation (Fig.

1,lanes 11 and 12).

Thus,

the

ability

ofthe CBP

complex

but notthe 24K-CBP to restorethe

cross-linking

of the 80-kDa

polypeptide

(prob-ably eIF-4B)

is consistent with

previous

results

demonstrat-ing

that the CBP

complex

is

absolutely

essential for the

cap-specific cross-linking

ofeIF-4B

(6),

and indicates that this

complex

is inactivatedin

poliovirus-infected cells,

con-sistent with earlier reports

(7, 9, 31).

In view of the fact that the 24K-CBP by itself does not restore

cap-specific cross-linking

upon addition toI-IF, we wantedtofind outwhytheamountof24K-CBP detected by

the

cross-linking

assay is

greatly

reduced in I-IF

compared

with thatdetected in U-IF. Itwas also

important

toaddress this

question

in

light

ofaprevious report from Hansen and Ehrenfeld

(12)

that theamount of 24K-CBP detectedbythe

cross-linking

assay in IFwas notreduced as aconsequence of

poliovirus

infection. Since the CBPcomplexseems tobe the factor that is

inactivated,

we tested theidea that

cross-linking

of the24K-CBPtomRNAis moreefficient when it is partof theCBP

complex

than when it is in the free form. If this is true, then a defect in the CBP complex might

indirectly

affect the

cross-linking

of the 24K-CBP. We

performed cross-linking

experimentswithpurified24K-CBP and CBP

complex (containing

an approximately equal amount of

24K-CBP). Figure

2A shows cross-linking of 24K-CBP which iscompletely

m7GDP

sensitive(lanes1and

2). It should be noted that ca. eightfold less 24K-CBP was used herecompared with the amount used in Fig. 1, lanes 9 and10, and that the exposure time is different. Thisexplains

the substantial difference in the amount of cross-linked 24K-CBP observed in the two cases. When the CBP com-plex was used, there was a much higher amount of m7GDP-sensitive cross-linked24K-CBP (comparelanes 3 and 4 with lanes 1 and 2 in Fig. 2A; note that the autoradiogram is overexposed to show the cross-linked polypeptide in lane 1). Figure 2B shows Coomassie blue staining of the samples

used for the cross-linking experiments. m7GDP

affinity-pu-rified rabbitreticulocyte CBP fromribosomeswas run on a sucrose gradient in 0.5 M KCI for resolution of the free 24K-CBP and the CBP complex which otherwise copurify

on the cap affinity column. The gel shows a section of the

gradient, withlane 1 towards thetop of the gradient. Lanes 1,2,and 3 inFig. 2representcontiguousgradient fractions. Lane 1 (Fig. 2B) shows a Coomassie blue stain of the material used for cross-linking analysis in Fig. 2A lanes 1 and 2, indicating the presence ofthe24K-CBPand a small amount of the 50-kDa polypeptide. Lane 3 (Fig. 2B) is a Coomassie blue stain of the CBP complex used for

cross-linking in Fig. 2A lanes 3 and 4, showing ca. twofold more 24K-CBP than in lane 1 (Fig. 2B) and the othermajor CBP

complex polypeptides (50 and -220 kDa [6]). The presence of the 50-kDa polypeptide (eIF-4A, in lane 1) sedimenting more slowlythan the purified CBPcomplex (lanes 2 and 3)

probably means that it tends to dissociate from the CBP

complex to a slight degree during centrifugation under

high-salt(0.5 M KCI) conditions. Itshould benoted that the

A

1 2 3 4

B 12 3

-P220

-elF-4A

24k-CBP-m7GDP:

-24k-CBP

Ew.-.

- +

FIG. 2. Autoradiograph of cross-linked 24K-CBP as the free polypeptide or as part ofthe CBP complex and Coomassie blue staining of the two forms of CBP. (A) Cross-linking, SDS-poly-acrylamide gel analysis, andautoradiography were asdescribed in the text. Lanes 1 and 2contained -0.25 ,ug of24K-CBP;lanes 3 and

4contained CBPcomplex containing-0.5 ,ugof 24K-CBP. m7GDP (0.67 mM) was included in lanes 2 and 4 as indicated below the figure. (B) SDS-polyacrylamide gel analysis of purified rabbit reti-culocyte CBP sedimented through a 0.5 M KCIsucrosegradient. Rabbit reticulocyte CBP purified from ribosomes was sedimented througha10to30% linear sucrosegradientinLCBcontaining0.5 M KCItoresolve the CBPcomplexfrom thefree 24K-CBP. Samples from across the gradientwere then resolved on a 10% SDS-poly-acrylamide gel and stained by Coomassie blue. A section ofthe gradientisshown;sedimentationwasfrom lefttoright (i.e.,lane 1 is toward the topofthegradient). Materialshown in lane 1(30,ll of thefraction fromthegradient)wasusedforcross-linking analysisin A,lanes 1 and 2. Materialshown in lane 3(30p.lof the fraction from thegradient)wasusedforcross-linking analysisinA, lanes 3 and 4.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.612.318.553.386.553.2]

amounts of 24K-CBP used in these experiments fall in the linear range for the cross-linking assay (data not shown). Thus, it is clear from these data that the cross-linking efficiency of the 24K-CBP is considerably higher when it is part of the CBPcomplex. This result most likely relates to the reduced level ofcross-linked24K-CBPobserved in I-IF, in which case the CBPcomplex isinactivated and indicates that the activity required to stimulate the cross-linking of 24K-CBP is impaired. We have also examined the transla-tional restoring activity of thefractionsshown in Fig. 2B and have found that it correlates with the presence of the 220-kDa polypeptide (data notshown), consistent with pre-vious observations (7, 31). Insummary,efficient cap-specific cross-linking of the 24K-CBP to mRNA and translational restoringactivity are bothdependent on the CBP complex. The results of Etchinson et al. (9), which indicate that poliovirus causes proteolysis of the 220-kDa polypeptide (P220) of the CBP complex, provide the first evidence of a particular structural defect in the CBP complex. Again, though, as in allpreviousattempts tocharacterize the defect in CBP caused by poliovirus, the approach was indirect. In an attempt to examine directly the abundance, structure,

and subcellular distribution of CBPs after poliovirus infec-tion, wepurifiedthemfromuninfectedorpoliovirus-infected

cells, using the m7GDPaffinity chromatography technique.

Hansen et al. (14) were able to detect the 24K-CBP by chemical cross-linking in the S100 fraction ofHeLa cells, and wehave purifiedhomogeneous 24K-CBPfrom the S100 fraction of rabbitreticulocytes.Consequently,we attempted topurify the 24K-CBPfrom theS100 fraction obtained from

equal amounts ofuninfected or infected cells; Fig. 3 shows anSDS-polyacrylamide gelanalysisofthepurified fractions.

Severalassays wereused toverify that theinfectedfractions used as starting material for the purification were actually infected, e.g., mRNA (capped versus naturally uncapped) specificity of the corresponding cell extracts in translation

and the presence ofviral antigens byimmunoblotting. Most

of the polypeptides present in the S100 fraction are not retained during passage through the m7GDP-coupled resin (e.g.,compare lanes 2 and 3 inFig. 3, which are the load and

flow-through, respectively, from uninfected material). Elu-tion with 100 ,uM GDP shows a single polypeptide of Mr

-60,000 that either has affinity for the GDP moiety of the affinity columnor, less likely, associates with the 24K-CBP

via aGDP-sensitive interaction (Fig. 3, lane 4). The amount

and size of this polypeptide are not affected by poliovirus

infection (compare lanes 4 and 8 in Fig. 3) Elution with 100 ,uM m7GDP yielded homogenous 24K-CBP (lane 5), which

comigrates with the 24K-CBP of rabbit reticulocytes (data notshown). Again, neithertheabundancenor the size of this

polypeptide was altered after poliovirus infection (compare

lanes 5and 9 in Fig. 3).

In light of some speculation that the 24K-CBP becomes phosphorylated during poliovirus infection (18), and to ex-amine the possibility that it undergoes some other kind of

covalent modification, we performed two-dimensional gel analysis (isoelectric focusing in the first dimension and

SDS-polyacrylamide gel electrophoresis in the second

di-mension) of the m7GDP affinity-purified protein from unin-fected and infected cells. Figure 4 shows Coomassie blue staining of the two-dimensional gels. It is clear that the

polypeptidesfromuninfectedand infectedcellscomigrate in both dimensions (Fig. 4C, mixture of24K-CBP from unin-fected andinfected cells) and that there is only one species

with a slightly acidic isoelectric point of -6.5. Thus, polio-virus infection has no effect on the size, abundance, or net

Mr

205k-1 205k-

16k-

97k-

68k-

43k-1 2 3 4 5

-sop

30k-UNINFECTED

6 78

2_C

,424k-CBP

#-9

424k-C3BP

INFECTED

FIG. 3. Purification of 24K-CBPfrom the S100 fraction of unin-fected and poliovirus-inunin-fected cells. m7GDP affinity chromatogra-phy was performed as described in the text. Purified fractions were resolved on a 10% SDS-polyacrylamide gel followed by Coomassie blue staining. Lane 1, 5 ,ul of molecular weight standards, 1 mg of protein per ml (Sigma); lane 2, -100 ,ug of material from uninfected cells loaded onto the m7GDP column; lane 3, -100 p.g of flow-through from uninfected material; lane 4, 40 ,ul (from a total of 1 ml) of GDP eluate from uninfected cells; lane 5, 50p.l (from a total of 500 ,ul)of m7GDP eluate obtained from uninfectedcells; lane 6, -200,ug of material from poliovirus-infected cells loaded onto the m'GDP column; lane 7,-200,ug of flow-through from infected cells; lane 8, 40 ,ul (from a total of 1 ml) of GDP eluate from infected cells; lane 9, 50 p.l (from a total of 500p.l) of m7GDP eluate from infected cells. Note that the total amount of protein loaded on theaffinity column was the same for uninfected and infected cells.

charge of the 24K-CBP isolated from the S100 fraction; neither does it impair its ability to bind to a capanalog, since it can be retained by and specifically elutedfrom the m7GDP

affinitycolumn. It was also important to examine the charge of the 24K-CBP associated with ribosomes in relation toits distribution between the free polypeptide and the CBP complex and hence in relation to the restoring activity. We performed these experiments with CBPisolatedfrom rabbit reticulocytes, since more manageable amounts of material are obtained from this source. Analysis of ribosomal 24K-CBP from rabbit reticulocytes (either asfree24K-CBPor as partof the CBP complex) showed the presence of two major isoelectric variants as previously reported (30). There were, however, noobvious differences in therelative abundanceof these forms when those associated with the CBP complex were compared with those isolated as the free 24-kDa polypeptide (data not shown). We have not performed this analysis with HeLa ribosomal 24K-CBP, and therefore we cannot exclude the possibility that it behaves differently.

However, since the activity that can restore translation of capped mRNAs in extracts from poliovirus-infected cells is associated with the CBP complex and is absent from the 24K-CBP isolated from ribosomes, it seems clear from the latter results that the restoring activity is not related to a particularisoelectric variant of the 24K-CBP.

Therabbitreticulocyte CBPcomplex is definedassuchby several criteria: (i) co-elution of the different polypeptides from them7GDP affinity column; (ii) co-elution and stability

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.612.322.562.74.285.2]

520 LEE, EDERY, AND SONENBERG

IEF

W

SDS

_A

4

24K

CBP

UNINFECTED

B

previously shown (6) andsomedegradation productofP220, identified as such by tryptic peptide mapping (data not shown) and the presence ofcommonantigenicdeterminants (9). Figure 5, lane 2 shows the GDP eluate obtained when material from uninfected cells was loaded on the m7GDP column. Again, as was the case forS100 fractions, a single polypeptide of Mr -60,000was eluted. This is presumably the same polypeptide as that obtained from the S100 frac-tions (Fig. 3, lane 4). Figure 5, lane 3 shows the m7GDP eluate obtained from uninfected cells. The 24-, 50-, and -220-kDa polypeptides comigrate with their reticulocyte counterparts,althoughthe -220-kDapolypeptideis asmear, presumably duetoproteolysis. In addition, therearebands of -60 and 70 kDa and other minor bands. The 60-kDa polypeptide isnotassociated with the CBPcomplexsince it can be completely removed by extensive washing of the affinity column with GDP before elution with m7GDP. The 70-kDa polypeptide is specifically eluted with m7GDP and may therefore correspond to the -70-kDapolypeptide pre-viously described in preparations of CBP II (11). These results show that the CBP complex from HeLa cells is

INFECTED

1

2

3

~

s-205k-*~

e.

FF---116k

-- -9k7

MIXING

FIG. 4. Two-dimensional gelanalysis of 24K-CBP from uninfect-edandpoliovirus-infected cells.Samples of m7GDPaffinity-purified 24K-CBP(from lanes5or9 inFig. 3)wereresolvedon two-dimen-sionalgels accordingtoO'Farrell (23), followed by Coomassie blue staining. (A)0.5 ,ugof24K-CBP from uninfected cells(U-24K-CBP); (B)0.5

p.g

of 24K-CBPfrompoliovirus-infected cells(I-24K-CBP); (C) mixture of-0.3 ,ugof U-24K-CBPand -0.3 ,ugofI-24K-CBP.

elF-4A- o_m

toseveralconventionalpurification stepsincluding gel filtra-tion; and(iii) cosedimentation ofcomponents of thepurified CBP complex in sucrose gradients containing 0.5 M KCI. Further indication that the 24-, 50-, and 220-kDa polypep-tides are complexed together comes from the facts that (i)

the 24-kDa polypeptide is the only polypeptide which, by itself, interacts with cap structures asassayed by chemical cross-linking (6), and (ii) the purified 50-kDa polypeptide (eIF-4A) doesnot bind to the m7GDP column(unpublished data). Thus, although the CBPcomplexhas notbeen rigor-ouslycharacterizedstoichiometricallyorbiophysically,there is good reasonto believe that it represents a homogeneous biological entity.

The CBPcomplexwas previously purifiedfrom the

high-salt washof rabbitreticulocyteribosomesbym7GDPaffinity chromatography (6, 9, 31),andweusedasimilarprotocolto purify it from HeLa cells, with the exception that the IF

were not fractionated with ammonium sulfate for technical convenience (see above). Figure 5 shows SDS-polyacryl-amidegel analysisof thepurified fractions. Lane 1 contains

asampleof rabbitreticulocyteCBPcomplex, showing major bands of 24(24K-CBP), 50 (eIF-4A), and -220kDa(P220)as

-3k30k

24K-CBP- C

0.C

o ro

E

I _4J I

o) a

0 r,(

L

4.-U

FIG. 5. Purification of CBP complex from the ribosomal salt

wash obtained from uninfectedand poliovirus-infectedcells.

Frac-tionswere purifiedon the m7GDPaffinity column,asdescribed in

the text,and resolvedon a10%SDS-polyacrylamide gel,followed

by silverstaining (20).Lane1, 0.5,ugof CBPcomplexfrom rabbit

reticulocyte ribosomes; lane2,40 ptl(fromatotal of4ml) of GDP

eluatefrom uninfectedcells;lane3, 30,u1 (from atotal of 1ml) of m7GDP eluate from uninfectedcells;lane4,40 p.1(fromatotal of4 ml)of GDP eluate frompoliovirus-infected cells;lane5,30 ,u1 (from

atotal of 1 ml)of m7GDP eluate frompoliovirus-infectedcells.

C

P220-Mr

4

5

" r .. ..7

C68kk

-43k

J.VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.612.81.274.73.419.2] [image:6.612.317.556.310.631.2]

structurally

very similar to the rabbit

reticulocyte

CBP

complex,

a result which accords with the

high

degree of conservation of

protein synthesis

initiation factors between

rabbits and humans(4).

Purification of CBP from the ribosomal high-salt wash

obtained from

poliovirus-infected

HeLa cells

yielded

dis-tinctly

differentresults

(Fig. 5,

lanes 4 and 5).The samples were runon adifferent gel from that shown inlanes 1 to 3;

[image:7.612.318.560.76.305.2]

the

corresponding

molecular

weights

are indicated in the

figure.

Lane 4showsthe GDPeluate obtainedfrom infected

cells, showing again

a -60-kDa

polypeptide.

The amount and sizeof this

polypeptide

again were not changed due to

poliovirus infection (compare lanes 4 and 2). There is also

staining

oneither side ofthe -60-kDa

polypeptide

that isan artifact ofthe silver

staining procedure

(20) and does not represent

purified polypeptides. Figure

5, lane 5 shows the

m7GDPeluateobtained from

poliovirus-infected

cells. There was no

change

in eithertheamount orthe sizeofthe24-kDa

polypeptide,

whereas the

remaining m7GDP-specific

bands were

distinctly

different.

First,

the -220- and 50-kDa

(elF-4A)

polypeptides

are almost

completely

absent.

Second,

therearetwonewbandsof-130kDathatarenotpresent in the

preparation

from uninfectedcells(compare lanes5 and3

in

Fig.

5). Although

it should be borne in mind that silver

staining

of

polypeptides

is not

necessarily quantitative,

it does appear that the amountofthe -130-kDa

polypeptides

is

significantly

less thanthe amountof24-kDa

polypeptide,

particularly

on amolar basis. The presence of 24K-CBPin the

m7GDP

eluate obtained from infected cells serves as a useful internal control and argues

strongly against

nonspe-cific loss ofthe other CBP

complex

polypeptides (50

and -220

kDa). Furthermore,

theobservationthat

homogeneous

24K-CBP is obtained

by m7GDP affinity purification

ofthe

postribosomal

supernatantfrominfected cells indicatesthat theCBP

complex

isnot

merely

redistributed inthe infected

cell such that it no

longer

associateswith ribosomes.

Thus,

a reasonable

interpretation

ofthese resultsis that theCBP

complex

is modified after

poliovirus

infection. The

signifi-canceof the -130-kDa

polypeptides

will be addressed later in

light

of the results shown in

Fig.

6andother data.

The

purification

of

large

amounts of the CBP complex fromHeLa cells isa somewhat cumbersome and

time-con-suming activity, yielding only

ca. 50 ,ug ofCBP

complex

from

1010 log-phase

HeLacells

(20

liters ofcells at 5 x 105 cells per

ml).

There are inaddition many steps between the

cell harvest and the m7GDP

affinity purification,

possibly

contributing

to artifactual

disintegration

ofthe native CBP

complex

as it exists in the cell.

Consequently,

we decided to label mock- and

poliovirus-infected

cells with

[35SJmethionine

and attempt to isolate the CBP

complex

by

a faster

protocol.

Cells were labeled for 6 h with

[35S]methionine.

At the end of this

time,

the cells were divided

equally

in two; halfwere infected with

poliovirus

and the other half were mock-infected. To monitor the

infectionwe

performed

a

mock-labeling experiment

in which

[35S]methionine

was added at 2.5 h

postinfection

to

previ-ously

unlabeled

cells,

in the presence and absenceof

polio-virus. The in vivo

labeling

pattern observed between 2.5 and 3 h

postinfection

(at which time the cell extracts were

prepared)

confirmedthat shut-off of cellular

protein

synthe-siswas

complete

and that

virus-specific proteins

were

being

synthesized (data

not

shown).

Figure

6 shows the results of the

purification

of

[35S]methionine-labeled

CBPfromuninfectedand

poliovirus-infected HeLa cells. Elution of the

m7GDP

columnwith100

,uM

GDP

yielded

a

single polypeptide

ofMr-60,000 (lane 1,

2 3 4 Mr 5 6 7 8

am

-220k-pw

_w

-a

-9 -P220

wa

-r

--W-60k

-50k

-24k

u u

-24k-CBP

u u

FIG. 6. Purification of in vivo-labeled CBP complex from unin-fected and poliovirus-infected HeLacells. Fractions werepurified on anm7GDPaffinitycolumnasdescribed in the text, resolvedona 10% SDS-polyacrylamide gel, and autoradiographed. Lane 1,40,ul (fromatotalof1 ml) of GDP eluate obtained from uninfected cells; lane2,40,u (from atotal of 1 ml) of GDP eluate from poliovirus-infectedcells; lane 3,50pI(fromatotalof 500pI)ofm7GDP eluate from uninfectedcells;lane4,50p.1(fromatotalof 500p1)ofm7GDP eluate from poliovirus-infected cells. Polyclonal antiserum against theCBPcomplexwasusedto probeextractsfrom uninfected and poliovirus-infected HeLa cells for P220 related antigens. S10 ex-tracts wereresolvedon a10% SDS-polyacrylamide gel, followed by Western blottingasdescribed inthe text. Thefigure shows immuno-reactivespeciesinlanes 5to9. Lanes5and 6, 150pugofprotein from differentS10extractsfrom uninfectedcells; lanes7and8,150p.g of protein from different S10extracts from poliovirus-infected cells; lane 9,5pugof rabbit reticulocyte CBP complex.

uninfected). This is presumably the same polypeptide ob-served when unlabeled materialwasusedfor the purification of 24K-CBP from the S100 fraction (e.g., Fig. 3, lane 4).

Again, the amount and size of this polypeptide were not

affectedby poliovirusinfection (compare lane 2withlane1). Lane 3 shows the m7GDP eluate obtained from uninfected material(includingthe 60Kpolypeptide whichisparticularly

abundant and isnotcompletelywashedoffduringtheelution

with GDP in this experiment). The material eluted has

polypeptides comigrating with the 24-, 50-, and 220-kDa

polypeptidesof rabbit reticulocyteCBP complex(indicated

by molecularweight to the right oflane 4). However, we havenoevidenceatpresenttoprovethat the 24-and 50-kDa

polypeptides are indeed the 24K-CBPand eIF-4A,

respec-tively. The other bands (Mr = 35,000 and 70,000, etc.) are either relatedto theCBPcomplex-since theyare

specific-allyeluted with m7GDP(compare lanes 3 and 1)-or, alter-natively, are other proteins which bind specifically to the column(thecriterion forspecificity beingelution with m7GDP andnotwithGDP).Since the loadontothecolumncontains totalsoluble cellproteinand alsoproteins solubilizedby0.5 MKCI and 0.5% NonidetP-40, it is

likely

thatatleastsome of the additional polypeptides are not related to the CBP

complex. For example, likely candidates are

putative

nu-clearCBPsas

previously reported

(24). Another

possibility

on November 10, 2019 by guest

http://jvi.asm.org/

is that the additional polypeptides are loosely associated

with the CBP complex but are lost during the purification protocols

previously

employed (6, 11, 31).Itshould benoted that the relative

labeling

intensities of the 24-, 50-, and -220-kDa polypeptides are not equal. This might be ac-countedfor in partbythe size of thepolypeptides (assuming an average methionine content for each

polypeptide)

but may also reflect differentratesofentry ofthe newly synthe-sized components into the CBP complex. In

addition,

the relative

labeling

intensities of the 24-, 50-, and 220-kDa

polypeptides

vary among different

preparations

from

unin-fected cells

(unpublished

data). Whenmaterial from poliovi-rus-infected cells was loaded onto the column, the m7GDP

eluate obtained was

distinctly

different from that obtained from uninfected material

(Fig.

6, lane 4). There was no

change

in either the abundance or the size ofthe 24- and 50-kDa

polypeptides,

whereas in contrast there was no detectable -220-kDa

polypeptide.

Instead, therewere

addi-tional bands ofMr -130,000 (indicatedby arrows)thatwere

completely

absent from the m7GDP eluate obtained from uninfected cells (compare lane 4 withlane 3 in Fig. 6). The

size and abundance of all the other bands present in the

m7GDP

eluate are also not affected by poliovirus infection (comparelane 4 with lane 3).

These results demonstrate that P220 is cleaved by a

putative

viral (orinduced) protease, but that thepresumed

cleavage products (-130-kDa polypeptides)

are stillretained andcanbe

specifically

eluted fromthe

m7GDP affinity

resin. This suggests that the presumed cleavage products remain associatedwitha

cap-binding

component(mostprobablythe

24K-CBP)

intheform ofamodified CBP complex. Itseems to usextremely

unlikely,

although

admittedly

notprecluded,

that the

cleavage

products derivedfromP220 would have a

cryptic m7GDP binding

site and thus bind directly to the

m7GDP affinity

resin. The difference in the amounts of the -130-kDa cleavage products obtained in Fig. 6 (lane 4) as

compared

with

Fig.

5 (lane 5) suggest that the putative modified CBP

complex

is not stable to the purification protocol

employed

fortheexperiments in Fig. 5. This may also

explain

theabsenceof the 50-kDapolypeptideinFig.5. Theexistenceof

proteolytic

cleavage

fragments

of P220 of

Mr

-130,000

accords withthe

original

observation of

Etch-ison et al. (9), who demonstrated the appearance of such

polypeptides

in crude HeLa cell extracts after

poliovirus

infection. In this case, the

cleavage

products observed are

thought

to berelated toP220oftheCBPcomplex byvirtue

ofcommon

antigenicity.

Wehave

recently

raised

polyclonal

antibodies to the

purified

sheep CBP complex which bind

strongly

tothe 24- and 220-kDa

polypeptides

of rabbit CBP

complex

(Fig.

6, lane 9).

Consequently,

we probed HeLa

cell extracts with this antiserum and obtained results very

similar tothose of Etchisonet al. (9). In

Fig.

6, lanes 5 and

6, differentextracts from uninfected

cells,

and in lanes 7 and

8,

different extracts from

poliovirus-infected cells,

were

probed

with theanti-CBPcomplexantiserum. The antiserum does not react with the HeLacell 24K-CBP (lanes 5to 8).

However,

the antiserum clearly reacts with the -220-kDa

polypeptide

presentinextractsfrom uninfected cells

(lanes

S and 6), and this antigen comigrates with P220 of the rabbit CBP

complex

(compare lane 5 with lane 9). There is no detectable 220-kDapolypeptide in extractsfrom

poliovirus-infected HeLa

cells,

whereas thereare putative degradation

products

Of

Mr

-130,000(Fig. 6, lanes7and8,indicatedby arrows). These latter bands comigrate with the cleavage products present in the putative modified CBP complex isolatedby

m7GDP

affinity chromatographyfrom

poliovirus-infected cells (compare lane 7 or 8 with lane 4), strongly

suggesting thatthey areidentical. DISCUSSION

Analysis of eucaryotic mRNA CBPs by the chemical cross-linking assay indicatesacomplex interactionprimarily

between polypeptides of 24, 50, and 80 kDa, ATP, and mRNA. This interaction is presumed to facilitate 40S ribosomal subunit attachment to cellular mRNA during translation and is very discretely prevented uponpoliovirus infection of HeLa cells, resulting in shut-off of cellular protein synthesis. Recently, Edery et al. (6) have demon-strated that the cap-specific mRNA-protein interaction ob-served between crude IF and mRNA can bereconstituted by using the CBP complex (containing eIF-4A as a subunit), eIF-4B, and mRNA. These factors appear to interact with mRNA in close concert, since the m7GDP-sensitive cross-linking of eIF-4A as part of the CBP complex is strictly dependent on eIF-4B and the cross-linking of eIF-4B is likewise dependent on the CBP complex. This idea accords with the simultaneous loss of the cross-linking ability of all thecap-specific polypeptides afterpoliovirus infection (19), againconsistent withaclosefunctionalrelationshipbetween them.Thus, the virally induced lesion in the CBP complex is probably sufficient by itselftopreventinteraction of eIF-4A and the 80-kDapolypeptide (probably eIF-4B) with the cap structureandconsequentlyblocks40Sribosomeattachment to cellular mRNAs. The CBP complex has activity which restores the specific cross-linking profile when added to IF

from infected cells. This activity is not present in the

24K-CBPand thuscopurifieswith thetranslationalrestoring activity, strongly suggesting that the two activities are identical. The fact that eIF-4A and eIF-4B are neither

structurally modified (5) norfunctionally impaired(16) after

poliovirus infection is consistent with ourobservation that the 80-kDa polypeptide (probably eIF-4B) is present and active in IF from infected cells, at least as assayed by chemical cross-linkingto mRNA in the presenceof exoge-nous CBPcomplex.

The 24K-CBP from infected cells, either as the free

polypeptide or as part of the putative modified CBP com-plex, can recognize the cap structure, as indicated by the fact that it canbe purified by m7GDPaffinity

chromatogra-phy (Fig. 3 and 5). However, the amount of cross-linked 24K-CBP in I-IFis considerablylower than that from U-IF. Inlight ofourfindingthat thecross-linkingof the 24K-CBP from rabbitreticulocytesisgreatly enhanced when it is part of the CBPcomplexascomparedwith the freepolypeptide,

it seems likely that the 24K-CBP in the putative modified CBPcomplexfrompoliovirus-infectedcells behaves like the free 24-kDapolypeptideintermsofcross-linkingtomRNA. This again points to asignificant role for P220 in mediating the interaction between the 24K-CBP and mRNA.

We have presented direct eivdence that the 24K-CBP is notstructurallymodified afterpoliovirusinfection. Further-more, the subcellular distribution of the 24K-CBP is not changed. In contrast, the native CBP complex cannot be

purifiedfrom anyfraction obtained frompoliovirus-infected

cells. These resultsdemonstratedirectlythatthe native CBP

complex is modified by poliovirus infection. The in vivo

labeling experimentsindicate thata modified CBPcomplex

exists in infected cellswhich contains theproteolytic cleav-ageproducts ofP220, and possibly eIF-4A. Similar results are obtained for purification of unlabeled ribosomal CBP from infectedcells, although in thiscaseeIF-4Aisdefinitely

on November 10, 2019 by guest

http://jvi.asm.org/

absent andthe amountof the-130-kDa

polypeptides

seems

significantly

reduced.

Thus,

theexact structure of the mod-ified CBP

complex

is uncertain. In other

experiments,

we have obtained the24K-CBPinfree form from ribosomalhigh

salt washofinfected

cells,

which suggests that the modified CBP

complex

is unstable. In

addition,

the eIF-4A compo-nent of the CBP

complex

is

apparently

not as

strongly

associated withthe CBP

complex

asthe other components, as indicated

by

the observation that a small amount of eIF-4A dissociates from the CBP

complex

under

high-salt

conditions

(Fig. 2B).

This

might

well

explain

the lack of eIF-4A in theribosomal CBP isolated frominfected

cells,

in

which caseP220is

cleaved, possibly resulting

indecreased

affinity

of eIF-4Afor other

complexed

components. In any event, theresults

presented

here suggest thatanintactP220 is essential for CBP

complex

function.

First,

efficient

cap-specific cross-linking

of the

24K-CBP, eIF-4A,

and eIF-4B is

dependent

on the CBP

complex

and does not occur after

poliovirus

infection.

Second, restoring activity

islikewise a propertyofthe CBP

complex

andcorrelates with the pres-enceof P220as

opposed

tothe24K-CBPoreIF-4A.Whether or not the association of the modified CBP

complex

with

ribosomes from infected cells reflects an involvement in translation of

poliovirus

RNA remains to be determined.

Indeed,

the mechanism

by

which

poliovirus

RNA initiates

translation is

something

ofamystery, both in terms ofany

possible

roleof

CBP(s) (modified

or

otherwise)

and

concern-ing

which structural features of the viral messenger allow efficienttranslation intheabsence ofthe cap structure.

The

question concerning

the

relationship

between the

virus-dependent

protease that cleaves P220 and the

poliovi-rus

replicative

cycle

now

challenges.

One

possibility

is that the

activity

that processes the

poliovirus

primary cleavage

products

(3C) is also

responsible

for

cleavage

of P220.

However,

this appears

unlikely,

in

light

ofour recent data thatantibodies directed

against poliovirus

protein

3C donot inhibit

cleavage

of P220invitro

(K.

A. W.

Lee,

Isaac

Edery,

R.

Hanecak,

E.

Wimmer,

and N.

Sonenberg,

submitted for

publication). Therefore,

there

might

be another viral prote-ase involved in this

cleavage

or induction of a cellular

function, possibly

one which is involved in

regulation

of

protein synthesis

in a broader sense. The

availability

of mutants derived from infectious cloned

poliovirus

DNA should aid in

approaching

theseproblems.

ACKNOWLEDGMENTS

We extendourthankstoChristiane BabinandSheelin Howard for assistancein preparationof viralRNA.

This researchwassupported bygrantsfrom the Medical Research Council and the NationalCancer Institute of CanadatoN.S. N.S. is therecipientofaTerryFoxCancerResearchScientist awardof the National Cancer Institute of Canada. K.A.W.L. and I.E. are

predoctoralfellows oftheCancer Research SocietyofMontreal.

LITERATURE CITED

1. Baltimore,D.1969. Thereplicationofpicornaviruses,p. 101-176. In H. B. Levy (ed.), The biochemistry of viruses. Marcel Dekker, Inc., NewYork.

2. Banerjee, A. K. 1980. 5'-Terminal cap structurein eucaryotic messengerribonucleic acids. Bacteriol.Rev. 44:175-205. 3. Banerjee,A.K.,and A.J.Shatkin.1970.Transcriptionin vitro

byreovirus-associated ribonucleicacid-dependent polymerase. J. Virol. 6:1-11.

4. Brown-Luedi,M.L.,L.J.Meyer,S.C.Milburn,P.M.-P.Yan, S. Corbett, and J. W. B. Hershey. 1982. Protein synthesis initiation factors from human HeLa cells and rabbit reticulo-cytesaresimilar:comparisonofproteinstructure,activities and immunologicalproperties. Biochemistry21:4202-4206.

5. Duncan, R., D. Etchison, andJ. W.B. Hershey. 1983. Protein synthesis initiation factors eIF-4A and eIF-4Bare notalteredby poliovirus infection of HeLa cells. J. Biol. Chem. 258:7236-7239. 6. Edery, I., M. Humbelin, A. Darveau, K. A. W. Lee, S.Milburn, J. W. B. Hershey, H. Trachsel, and N. Sonenberg. 1983. Involve-mentof eukaryotic initiation factor 4A in the cap recognition process. J. Biol. Chem.258:11398-11403.

7. Edery, I., K. A. W. Lee, and N. Sonenberg. 1984. Functional characterizationofeukaryotic mRNA cap binding protein

com-plex: effects on translation of capped and naturally uncapped RNAs. Biochemistry 23:2456-2462.

8. Ehrenfeld, E. 1982. Poliovirus-induced inhibition of host cell protein synthesis.Cell 28:435-436.

9. Etchison, D., S. C. Milburn, I. Edery, N. Sonenberg, and J. W. B. Hershey. 1982. Inhibition of HeLa cellprotein synthe-sisfollowingpoliovirus infection correlates with the proteolysis of a 220,000-dalton polypeptide associated with eucaryotic initiation factor 3 and acapbinding protein complex. J. Biol. Chem. 257:14806-14810.

10. Grifo, J. A., S. M. Tahara, J. P. Leis, M. A. Morgan, A. J. Shatkin, and W. C. Merrick. 1982. Characterization of eukaryo-tic initiation factor4A, a protein involved in theATP-dependent bindingofglobinmRNA.J. Biol. Chem. 257:5246-5252. 11. Grifo, J. A., S. M. Tahara, M. A. Morgan, A.J. Shatkin, and

W. C. Merrick. 1983. New initiator factoractivity required for globinmRNA translation. J. Biol. Chem.258:5804-5810. 12. Hansen, J.,and E.Ehrenfeld.1981.Presence of the cap-binding

protein in initiation factor preparations frompoliovirus-infected HeLa cells. J. Virol.38:438-445.

13. Hansen, J., D. Etchison, J. W. B. Hershey, and E. Ehrenfeld. 1982. Association ofcap-binding proteinwitheucaryotic initia-tion factor 3in initiation factorpreparationsfromuninfected and poliovirus-infectedHeLa cells. J. Virol. 42:200-207.

14. Hansen, J. L., D. 0. Etchison, J. W. B. Hershey, and E. Ehrenfeld. 1982. Localization ofcap-bindingprotein in subcel-lularfractions of HeLacells. Mol. Cell. Biol. 2:1639-1643. 15. Helentjaris, T., and E. Ehrenfeld. 1978. Control of protein

synthesis in extracts from poliovirus-infected cells. I. mRNA discriminationby crude initiationfactors. J. Virol. 26:510-521. 16. Helentjaris, T.,E.Ehrenfeld,M. L.Brown-Luedi, andJ. W. B. Hershey. 1979. Alterations in initiation factor activity from poliovirus-infectedHeLa cells.J. Biol. Chem. 254:10973-10978. 17. Hewlett,M.J., J.K.Rose,and D.Baltimore.1976. 5'-Terminal

structureofpoliovirus polyribosomalRNA ispUp. Proc.Natl. Acad. Sci. U.S.A. 73:327-330.

18. James, A. L., and D. R. Tershak. 1981. Protein phosphoryla-tionsinpoliovirus infected cells. Can. J. Microbiol. 27:28-35. 19. Lee, K. A. W., and N. Sonenberg. 1982. Inactivation ofcap

binding proteins accompanies the shut off of host protein synthesis by poliovirus. Proc. Natl. Acad. Sci. U.S.A. 79:3447-3451.

20. Merril, C. R., D. Goldman, and M. L.Van Keuren. 1984. Gel proteinstains: silver stain. Methods Enzymol. 104:441-447. 21. Muthukrishnan, S., M. Morgan, A. K. Banerjee, and A. J.

Shatkin. 1976.Influenceof5'terminalm7G and2'-O-methylated residues on messengerribonucleic acid binding to ribosomes. Biochemistry 15:5761-5768.

22. Nomoto, A., Y. F. Lee,and E. Wimmer. 1976. The 5' endof poliovirus mRNAisnotcapped with m7G(5')ppp(5')Np. Proc. Natl.Acad. Sci. U.S.A.73:375-380.

23. O'Farrell, P. H. 1975. High resolution two-dimensional elec-trophoresisofproteins.J. Biol. Chem.250:4007-4021. 24. Patzelt, E., D. Blaas, and E. Kuechler. 1983. CAP binding

proteins associated with the nucleus. Nucleic Acids Res. 11:5821-5835.

25. Rose, J. K., H. Trachsel, K. Leong, and D. Baltimore. 1978. Inhibition of translationby poliovirus: inactivationofaspecific initiation factor. Proc. Natl.Acad.Sci. U.S.A. 75:2732-2736. 26. Schreier,M.H.,and T.Staehelin. 1973.Initiation ofmammalian

protein synthesis: the importance of ribosome and initiation factorqualityfor theefficiencyof in vitro systems. J.Mol.Biol. 73:329-349.

27. Shatkin, A. J. 1976. Capping of eukaryotic mRNA. Cell

on November 10, 2019 by guest

http://jvi.asm.org/

9:645-653.

28. Sonenberg, N. 1981. ATP/Mg2 -dependent cross-linking ofcap

binding proteins to the 5' end of eucaryotic mRNA. Nucleic Acids Res. 9:1643-1656.

29. Sonenberg, N., M. A. Morgan, W. C. Merrick, and A. J. Shatkin. 1978. Apolypeptide ineukaryoticinitiation factors that crosslinks specifically to the 5'-terminal cap in mRNA. Proc.

Natl. Acad. Sci. U.S.A.75:4843-4847.

30. Sonenberg, N., K. M. Rupprecht, S. M. Hecht, and A. J. Shatkin. 1979.EukaryoticmRNAcapbindingprotein:

purifica-tion by affinitychromatographyonsepharose coupledm7GDP.

Proc. Natl. Acad. Sci. U.S.A.76:4345-4349.

31. Tahara, S., M. A. Morgan, and A. J. Shatkin. 1981.Twoforms ofpurified m7G-cap binding protein with different effects on

cappedmRNAtranslation inextracts ofuninfected and polio-virus-infected HeLa cells. J. Biol. Chem. 256:7691-7694. 32. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic

transfer ofproteins from polyacrylamide gelsto nitrocellulose sheets: procedures and some applications. Proc. Natl. Acad.

Sci. U.S.A. 76:4350-4354.

33. Trachsel,H.,N.Sonenberg,A.J.Shatkin,J. K. Rose, K.Leong, J. E. Bergmann,J.Gordon, andD.Baltimore.1980. Purification ofafactor thatrestores translation of VSV mRNA inextracts

from poliovirus-infected HeLa cells. Proc. Natl. Acad. Sci. U.S.A. 77:770-774.

34. Vaitukaitis, J. L.1981. Production of anti-sera with small doses of immunogen: multiple intradermal injections. Methods Enzymol. 73:46-52.

on November 10, 2019 by guest

http://jvi.asm.org/