Structural and functional characterization of the poliovirus replication complex.

0022-538X/92/052740-08$02.00/0

Copyright ©3 1992, AmericanSocietyfor Microbiology

Structural and Functional

Characterization

of the

Poliovirus

Replication Complex

KURTBIENZ,* DENISE EGGER, THOMAS PFISTER, AND MONICA TROXLER InstituteforMedical Microbiology, UniversityofBasel, Petersplatz 10, CH-4003Basel, Switzerland

Received 18November1991/Accepted 10 February 1992

Two populations of membrane-bound replication complexes were isolated from poliovirus-infected HEp-2

cells bysucrosegradient centrifugation. Thetwofractions show similar ultrastructural features: the replication complex is enclosed ina rosettelike shell of virus-induced vesicles and contains a verytightlypacked second

membrane system (compact membranes). The vesicular fraction, which bands in 30% sucrose, contains

replicative intermediate (RI) and 36S RNA. The fraction banding in 45%sucrosecontains only minuteamounts

of RI and contains mainly 36S RNA, two-thirds of which is encapsidated. In vitro, the two fractions show similarRNAsynthesizing capacities and produce 36Splus-strandRNA. Dissolving themembranes within and around synthetically active replication complexes with sodium deoxycholate abolishes the completion of36S RNAbutstill allows elongationinthe RI. Ourfindingssuggestanarchitecture of thereplication complex that has thenascentplus strandsonthe RIenclosed in thecompactmembranes and the replication forks wrapped additionally inprotein. Plus-strand RNAcanbe localizedbyinsituhybridization withabiotinylatedriboprobe between thereplication complex and therosetteof the virus-induced vesicles. Itwas found that theprogeny RNAstrands aresetfreesoon aftercompletion from the replication complex at the sites where thecompact membraneswithin thereplication complexare inclosecontactwith thesurrounding virus-induced vesicles.

The replication of poliovirus RNA is carried out by the primer-dependent viral polymerase 3DPo' and proceeds asymmetrically in two distinct steps. First, the input plus-strand(genomic) RNA iscopied into aminusstrand, which leads to the formation of the double-stranded replicative form(RF) (1, 24). This reaction canbe studied invitro ina reconstituted system (28) and does not depend on any

cellular component or structure. The next step, however, whichproducesprogenyRNA strands ofpositive polarity, is fully dependent on specialized cellular membranous struc-tures(25). Itstartsfromthe RF andproceedsin thepartially double-stranded replicative intermediate (RI) (11), which is confined in the viral replication complex. The replication complexwasidentifiedby electron microscopy(EM) (3) and foundtobesurrounded by and stretchedoutinsidearosette

of virus-induced cytoplasmic vesicles (6). The structural configuration of the replication complex is essential for ongoingplus-strand RNA synthesis and is maintained by the membranes of the virus-induced vesicles in concert with viral protein 2C (6).

The strict dependence of the functional replication com-plex on cellular membranous structures has hitherto pre-cluded the use of truly reconstituted in vitro transcription systemstostudy viralplus-strandRNAsynthesis. By using appropriate subcellular fractions of infected cells, however, 36Sgenomic RNAcan be produced in vitro (6, 10, 11, 24, 25), and thus at least some of the requirements for this synthetic activitycan be investigated.

In thepresentstudy,weusedsuchsubcellular fractionsto

elucidatetherolesof differentcomponentsof thereplication complex and the surrounding vesicles in viral plus-strand RNA synthesis. By EM immunocytochemistry and in situ hybridization, we visualized the spatial relation of viral macromolecules involved in RNAsynthesis withrespectto

each otherandtothecellular (yet virus-induced)structures.

* Correspondingauthor.

This allowed us to draw conclusions about the functional interaction of the replication complex with its associated membranes. In addition to the known virus-induced vesi-cles, we could identify a second membrane system within the replication complex. This membrane system is very tightly packed and appearsto interact withthe surrounding virus-induced vesicles in the last stepsofcompletion of the

36S progeny RNA. We also found that elongation of

plus-strand RNA at the RI proceeds in the absence of mem-branes, whereas completion and liberation of36S RNAare dependenton membranes.

MATERIALS AND METHODS

Cells and virus and in vivo labeling ofviral RNA. HEp-2 cells and poliovirus type 1 (Mahoney) were grown in sus-pension cultures. Themultiplicity ofinfection for the exper-iments was 30 PFU per cell. To label viral RNA, 5 p.gof actinomycin D (Merck Sharp & Dohme, Rahway, N.J.)per mlwas addedtoinfected cells 30min before the addition of 50 p.Ci of [5-3H]uridine (Amersham, Amersham, United Kingdom) perml. The labelwas added at 2.75 h postinfec-tion and left until the cellswere harvested at 4 h postinfec-tion.

Isolation and in vitro RNA synthesis of replication

com-plexes attached to virus-induced vesicles. Preparation of cy-toplasmic extracts by Dounce homogenization and low-speed centrifugation and isolation of the vesicle-attached replication complexes were described previously (6). In short, cytoplasmic extracts were centrifuged onto a double

sucrose cushion (30% sucrose layered over45% sucrose in

reticulocyte standard buffer) for 1 h at 55,000 rpm in an

SW55 rotor. Thevesicular fractions synthesizedviral RNA whenintroduced into acell-free transcription system (6, 24) containing all four nucleotides and an ATP-generating

sys-tem. To monitor in vitro RNA synthesis, 50 p.Ci of [5,6-3H]UTP (Amersham) per40 p.l was added tothe transcrip-tion system.

2740

on November 9, 2019 by guest

http://jvi.asm.org/

In some experiments,membranes weredissolved by treat-ing the vesicular fraction with 0.5% sodium deoxycholate

(DOC; Difco) for 30 minon ice (5). To test the accessibility of viral RNA, the nativevesicular fraction was treatedwith 100 ,ugof bovine pancreas RNase A(Boehringer Mannheim) perml for 15 min at25°C.

Isolation and characterization of viral RNA. In vivo- or in vitro-synthesized viral RNA was isolated by the

phenol-chloroform-isoamyl alcohol method (23), precipitated with ethanol, redissolved in water, adjusted to 10 mM Tris-HCl (pH

7.4)-i

mM EDTA-2 mM dithiothreitol-1 U of RNase inhibitor from human placenta (Boehringer Mannheim) per

p.l,

and loaded on 5 to 20% sucrose gradients in 10 mM Tris-HCI (pH

7.4)-i

mM EDTA. Thegradients were centri-fuged for 2.5 h at50,000 rpm in an SW55 rotorand fraction-ated from the bottom, and theradioactivitywasmeasured in aliquid scintillationcounter.Sedimentationcoefficient stan-dardsincluded3H-labeled poliovirus36S RNAisolated from

purified virions (7)and Escherichia coli 16S and 23S rRNA obtained from Boehringer Mannheim.

Biotinylated RNA probes for in situ hybridization. For in situ hybridization on the negatively stained preparationsof vesicles and replication complexes, the RNAprobe utilized for sectioned material was used (26). In short, from a plasmid (pVR106 [22]) containing full-length poliovirus

cDNA, nucleotides 670 to 2099 were excised with BamHI,

cloned intoapGEM-3vector(PromegaBiotec)between the SP6 and T7 promoters, and linearized with SspI at

position

1639 of the insert. The sequence between nucleotides 1639 and 2099 was transcribed in vitro into RNA of negative polarity with 50 U of T7 RNApolymerase (BRL)and 1 mM biotin-11-UTP (Enzo).

MAb. Monoclonal antibodies

(MAb)

against proteins

of the P2 genomic region of

poliovirus

were characterized previously (3, 19). Theyrecognize an epitope in protein 2C and its precursors 2BC and P2. They will be referred to hereafter as 2C-MAb. The

hybridoma

culture supernatants wereused at a 1:2 dilution for

immunocytochemistry.

EM. Specimens for EM histochemistry, immunocyto-chemistry(IEM), and in situ

hybridization

were

prepared

as described previously (6) by adsorbing the vesicle fractions onto nitrocellulose

(Parlodion)-coated

EM

grids.

The prep-arations were negatively stained with 1%

phosphotungstic

acid, pH 7.0, before being viewed in a Siemens 102 EM operated at 100or 125 kV.

(i)Detection of viralP2 proteins byIEM. For

immunocyto-chemical labeling,

nonspecific binding

siteson the prepara-tions were blocked with normal goat serum and bovine serum albumin. Next, the

grids

were incubated with 2C-MAb and then with goat anti-mouse antibodies

coupled

to 5-nm colloidal gold (Janssen

Pharmaceutica)

as described

previously (6).

(ii) Detection ofRNAwithRNase-gold.For RNAdetection withRNase-gold(2),the same types of

specimen

and block-ing reaction utilized for IEMwere used. To visualize

RNA,

gridswere incubatedon RNase

coupled

to 15-nm

gold

(E-Y

Laboratories, SanMateo,Calif.).Carewastakento

keep

all solutions as RNase free as

possible.

(iii) Doublelabelingfor RNAand P2 proteindetection. The RNase-goldand the 2C-MAbwere mixedand

applied

tothe

preparation. Afterbeingwashed, the

grid

was treated with 5-nm gold-labeled anti-mouse

antibody

and

negatively

stained asdescribedabove.

(iv) EM in situ hybridization. The vesicle fractions de-scribed abovewerealsoused for in situ

hybridization,

which was performed essentially as described

previously

for

ul-dpm 45% ves. 40000 20000 dpm 30% ves. *20000 *10000 30 fractions

FIG. 1. RNAcontent of vesicular (ves.)fractions isolated from poliovirus-infected HEp-2 cells. The cells were labeled with

[3H]

uridine from 2.75 to 4 h postinfection, and RNA ofthe vesicular fractions was analyzed, afterphenol extraction, by centrifugation

through5to20% sucrosegradients.The30% vesicularfraction(A) contains 36S RNA and RI sedimenting at 27 to 30S. The 45% vesicular fraction (A) contains 36S RNA and a small shoulder (fractions20 to22)ofpresumably RI RNA.

trathin sections

(26).

Some modifications

proved

to be

es-sential to retain ultrastructural

preservation

of the

native,

and thus rather

fragile,

vesicles. After in vitro

transcription,

thevesicular fractions werewashed once

by

centrifugation

and

resuspended

in

phosphate-buffered

saline

(PBS).

UV

cross-linking

minimized the loss of RNA in the

subsequent

steps. The material was then adsorbed to EM

grids,

and

hybridization

was

performed

for 2 h at

58°C by

floating

the

grids

on

10-,ul drops

of a

hybridization

solution

containing

4x SSC

(lx

SSC is 0.15 M NaCl

plus

0.015 M sodium

citrate),

13.5%

formamide,

and the

biotinylated riboprobe.

Afterthe

grids

werewashed with

PBS,

biotin detectionwas

performed

immunocytochemically

with anti-biotin antibod-ies

coupled

to 6-nm

gold

(AURION, Wageningen,

The

Netherlands).

As a

negative control,

an unrelated

biotiny-lated RNA

probe

(vaccinia

virus

thymidine kinase)

was

used.

RESULTS

Content ofviral RNA

species

intwovesicle

populations.

The

poliovirus

replication complex,

attached tovesicular

mem-branes,

isfound intwodistinct

populations

of vesicles

(6);

it bands in 30 or 45% sucrose and is referred to asthe 30 or

45%

fraction,

respectively.

We

investigated

possible

differ-encesincontentof viral RNA

by

phenol

extraction followed

by

sucrose

gradient

centrifugation

of in vivo

[3H]uridine-labeled RNA.

Figure

1shows that the30% fraction contains 36S

genomic

RNAinadditiontoRI

sedimenting

at27to30S. The 45% fractions contained

largely

36S

genomic

RNAand asmallamountofmore

slowly

sedimenting

RNA of

presum-ably

RI

configuration.

RNase treatmentafter

phenol

extraction indicated thatin both vesicle

preparations,

no double-stranded RNA was

detectable

(not

shown).

If the vesicle

preparations

were

treatedwith RNase Abefore

phenol

extraction

(Fig. 2),

the 36SRNA inthe30% fractionwasfoundtobe

fully

accessible to the enzyme, whereas the RI was

protected.

In the 45%

fraction,

abouttwo-thirds of the 36S RNAwasfoundtobe

protected

against

RNase

(compare

the

corresponding

36S

on November 9, 2019 by guest

http://jvi.asm.org/

[image:2.612.315.552.77.235.2]

dpm

36 - 23 -16S 40000

-30000

20000

-10000

0-0 1 0 20 30

fractions FIG. 2. RNase treatment ofa30%vesicular fraction(A)digests practically all of the 36S RNA but leaves the RI intact. RNase treatmentofa45%vesicular fraction (A)digestsabout one-third of the 36S RNA, leaving the remainder intact because it is encapsi-dated (data not shown). RNA was analyzed as described in the legend toFig. 1.

peaks in Fig. 1 and 2). The protected RNA of the native, non-phenol-extracted45%fractionsedimentedat

150S

(data

not shown) and thus wasencapsidated invirions.

Transcriptional activity of the 30and45%fractionsinvitro. The amount and type of RNA synthesized in vitro were monitored for both vesicular preparations. The 30% frac-tions incorporated [3H]UTP at first into RI and gradually

also into 36S RNA (Fig. 3). The 45% fractions, which contained predominantly 36S RNA, showed kinetics of RI and36S RNAsynthesis in vitro (Fig. 4) similar to those of the30% fraction. Afterapproximately60min, both systems became exhausted. The invitro-synthesized36S RNA of the 30% fractionaswellasthatof the45%fractionwasfoundto befullysensitivetoRNase,whereas theRI ofboth fractions wasprotected (not shown). Thus, during invitro

transcrip-tion, both fractions react to RNase treatment in a manner similarto that offreshlyisolated 30% fraction.

The combined data from Fig. 1 to 4 are interpreted to

0 10 20 30

fractions FIG. 3. Invitro RNAsynthesis ofa30% vesicularfraction for15 min (*), 30 min (-), 45 min (A), and 60 min (-). [3H]UTP is incorporatedfirst into RI andgradually also into 36S RNA. RNA

wasanalyzedasdescribed inthe legendtoFig. 1.

dpm

40000 36S

30000-

20000-10000

0

0 10 20 30

fractions FIG. 4. Invitro RNAsynthesis ofa45% vesicular fractionfor15 min(O),30 min(O),45 min(A), and 60min(0). Thekinetics and pattern of plus-strand RNA synthesis are very similar to those obtained with30%fractions. RNA wasanalyzedasdescribedin the legendtoFig. 1.

meanthat the 30% fraction, since it contains RI,synthesizes

RNA in vivo and continues the synthesis in vitro, whereas the 45% fraction seems to restart its plus-strand RNA

synthesisin vitro.Whether thishappens aselongation, e.g., on a small amountof RI (trailing shoulder in Fig. 1) or as initiationon RFor RIcannotbe determinedatpresent(see Discussion).

Ultrastructural aspects of the 30 and 45% vesicles. As statedearlier (6),the poliovirus replication complex is func-tionalonly if, bymeans of viral protein 2C, it is associated with the membranes of the virus-induced vesicles. In the currentinvestigation, we found that the virus-induced

vesi-cies

are not the only membranesystem associated with the replication complex. A second type ofmembrane was ob-served in thereplication complex of bothvesicularfractions

(Fig. 5).These membranes are tightly packed and appear as small (approximately 50-nm), clustered, vesiclelike bodies

(compact membranes). Like the membranes of the virus-induced vesicles, these compact membranes were labeled with2C-MAb(Fig.5). The label was foundpredominantlyat theborder of thereplication complexatthesiteswhere there is intimate contact between the compact membranes of the replication complex and the membranes of the virus-induced vesicles.

The ultrastructural aspects of the compact membranes within the replication complex are the same in the 30 and 45% fractions and also do not change during in vitro tran-scription. In ultrathin sections of poliovirus-infected cells, the compact membranes appear tightly coiled and do not have a clearly defined shape (data not shown).

DOCtreatmentof vesicularfractions. Totestwhetherthe replicationcomplex-associated membranes described above are involved in the observedprotection of the RI, afreshly isolated, RI-containing 30% fraction was treatedwith DOC for 30 min at 0°C. This rendered the RI partially RNase susceptible,asthe RIcould now be convertedbyRNaseinto an 18S core (10) (Fig. 6). After phenol extraction, the 18S corewasRNase sensitive(data notshown), which indicates that this 18Sstructure is derived from theRI (andnot from the RF) and that it is enclosed inpresumably proteinaceous material.

Together with the findings illustrated in Fig. 2, these

on November 9, 2019 by guest

http://jvi.asm.org/

[image:3.612.69.309.77.242.2] [image:3.612.320.564.78.239.2] [image:3.612.68.308.530.712.2]

~~~U1

FIG. 5. Electronmicrographofapoliovirusreplicationcomplexsurroundedbyvirus-induced vesicles(V)andcontainingasecond, compact

membranesystem (arrowheads). Immunocytochemical labelingwith2C-MLAb and 5-nm colloidal gold shows thatboth membrane types are

labeled. Themicrographshowsa45% vesicular fraction. Nomorphologicaldifferences from the 30% fractions could be found. Bar,100 nm.

findings indicate that themature36S RNA is located outside of thereplication complex,whereas thenascentplusstrands of the RI are sequestered by membranes. DOC treatment converts the replication complexinto amembrane-deficient

dpm 36 - 23- 16S

200 -_ _ _ -

-20000

t

U

0 10 20 30

fractions FIG. 6. RNasetreatmentofa30%vesicularfractiondigests only 36S RNA and leaves the RI intact (A) (Fig. 2). The RI can be digestedto an 18S coreonly afterDOC treatment (-), indicating that membranes protect thegrowing plus strands. RNAwas

ana-lyzedas described in thelegendtoFig. 1.

matrix containing the RIwith protruding, and thus

RNase-sensitive, nascentRNA strands.

Since the 30 and45% fractionsare

morphologically

and in vitro alsofunctionally identical, thefunction(s) of the

repli-cation complex-associated membranes during in vitro viral RNA synthesis was investigated with 45% vesicular frac-tions.

DOCwasaddedtoan invitrotranscriptionsystemafter 30 minofongoing in vitroRNAsynthesis. Duringfurther RNA

synthesis, the amount ofradioactivity increased in the RI,

whereas theamountof36SRNAremainedstableatthe level reached when DOC was added (Fig. 7). Thus, it can be concluded that themembranesof the

transcriptionally

active

replication complexes not only sequester the nascent plus

strands but alsoareinvolved in their detachment in

becom-ing36S RNA.

Ultrastructural analysis of

replication complexes

after treatmentwithDOCshows that thevirus-inducedvesiclesas well as the compact vesiclelike structures within the

repli-cation complexhave completelydisappeared

(Fig. 8).

This indicates that these compact structures indeed consist of membranes. The membrane-deficient matrix of the

replica-tion complexstillcontains P2protein atits

periphery.

Role of membranes and location ofprogenyvirus RNA in thereplicationcomplexasrevealedbyRNase-gold histochem-istryand insituhybridization. The observations

presented

so far indicate that the replication complex is a membrane-containing structure that encloses the RI

tightly

and is

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.138.470.79.398.2] [image:4.612.54.296.522.683.2]

dpm

36 23-16S

30000 - l l

20000-10000 L

0'

0 10 20 30

fractions FIG. 7. Invitro RNA synthesis of a45% vesicular fraction was allowed to proceed for 30 min, and RI and 36S RNA were made (A). To a parallel tube (O), DOC was added after 30 min of in vitro synthesis, and the reaction was allowed to proceed for another 15 min; only RIcontinued to incorporate [3H]UTP,whereas radioac-tivity did not increase in the 36S moiety. RNA was analyzed as described in thelegend to Fig. 1.

surrounded by a rosette of virus-induced vesicles. On the basis of the ultrastructural data(Fig. 5), it seemed to us that the resistance of the nascent plus-strand RNA against RNase reported above might be mediated not by the rosettes of large, virus-induced vesicles (which do not seemtobe in an extremely tight arrangement) but by the densely packed compact membranes within thereplication complex.

Toidentifythe membrane systemprotectingthe RI andat thesame timeto obtainmore information about the topog-raphy of thereplicationcomplex,weexploredthe location of

the 36S RNA after in vitro RNA synthesis, adopting the following rationale. Since only the 36S RNA and not the RI was found to beaccessible to RNase, only the 36S RNA can be located by histochemistry and in situ hybridization. The observedlocation of the 36S RNA would then be indicative of thetightness of the two membrane systems.

InFig. 9, RNA in general was identified by a histochem-ical reaction using RNase coupled to colloidal gold. In Fig. 10,viral 36S plus-strand RNA was localized specifically by in situ hybridization with a biotinylated riboprobe comple-mentary to a partof theP1 genomic region. Both reactions clearly show their respective targets located within the rosetteof thevirus-induced vesicles, thus demonstrating the rather loose and in any casepermeable configuration of the vesicular rosette and suggesting that the central compact membranes are the protecting agents of the RI. It is note-worthythat the 36S RNA is set free at several distinct sites at the periphery of the replication complex. This can be explained by assuming a topography of the viral RNA replication in which the template moves through the repli-cation complex and has the progeny RNA strands fixed at defined sites.

DISCUSSION

From poliovirus-infected cells, two populations of mem-brane-bound replication complexes can be isolated. One population bands in 30% sucrose and actively synthesizes

36S plus-strand RNA in vivo and in vitro. The fraction banding in 45% sucrose contains capsid precursors and virions (20) and besides encapsidated and free 36S RNA contains little, if any, RI. In vivo, the 45% fraction is therefore considered to be transcriptionally silent ("burnt out") but contains progeny RNA in the process ofbecoming

encapsidated. In vitro, it is capable of resuming RNA

I ..

*,

.

.4~f

V

-kN .

.0-FIG. 8. ElectronmicrographofaDOC-treatedreplicationcomplex(45% fraction). The membranesof the virus-induced vesiclesaswell

asthe compact membranes within thereplicationcomplexaredissolved. Theimmunocytochemical label with 2C-MAb isatthe surface of thecorestructureof the replicationcomplex. Bar, 100nm.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:5.612.65.306.78.239.2] [image:5.612.125.514.456.696.2]

[image:6.612.56.290.75.332.2]

FIG. 9. Avesicle (V)-bound replication complex (RC) as in Fig. 5 but doubly labeled with 2C-MAb (small gold grains) and with RNase-gold (large gold grains). The 36S RNA is located between the virus-induced vesicles and the replication complex. Bar, 100 nm.

synthesis and producing 36S RNA (Fig. 4). It is not clear whether in vitro the plus-strand RNA is elongated only on a small amountof RI or whether there is true initiation, which could take place either on preexisting RI or on RF, which could already be present in small quantities in vivo orcould have been freshly synthesized in vitro.

The two typesof viral replication complexes show similar ultrastructural features. They are enclosed in a rosettelike shell of virus-induced vesicles, which are attached to the replication complex by means of viral protein 2C (6). In the present investigation, we identified a second, very compact membrane system within the replication complex itself (Fig. 5). It encloses tightly the RI, which in concert with viral proteins represents the actual plus-strand RNA-synthesizing machinery. This was shown by our DOC experiments: removal of all membranes with DOC leaves the core of the replication complex, which still incorporates [3H]UTP. The finding that under these conditions [3H]UTP was incorpo-rated only into RI and not into 36S RNA indicates that elongation takes place but that no completed 36S RNA is released. Theelongation proceeds forquite some time, albeit at a reduced rate compared with that of non-DOC-treated replication complexes or intact cells.

IfDOC-treated cores of replication complexes devoid of membranes are treated with RNase A, the nascent plus strands at the RI areexposed to the action of the enzyme, whereas the partially double-stranded template with the replication forks remains protected from the RNase by a substance removable by phenol. RNase treatment of non-DOC-treated, transcriptionally active replication complexes

FIG. 10. Avesicle(V)-bound replication complex(RC)(45%fraction)onwhich insituhybridizationwithabiotinylatedRNAprobewas performed.Theprobe is visualized with6-nmgold-coupledanti-biotin antibodies. Thehybridization signal, localizingthe 36SRNA,is found onthesurface of thereplication complexbut within therosetteofthe virus-inducedvesicles.(Inset)Highermagnification.Thehybridization signalis found wherethe compact membranes(CM)of thereplication complexarein closecontactwith thevirus-induced vesicles.Bars,100 nm.

.:-...0.4.

.. -low

.,P:

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.132.472.396.676.2]

digests only36SRNA whileleavingthe entire RI intact. This means that freshly completed 36SRNA is rapidly set free. Since the completed 36S RNA is found by RNase-gold as well as by in situ hybridization within the rosette of virus-inducedvesicles but outside thereplication complexper se, it is possible that the compact membranes within the repli-cation complex have to interact with the virus-induced vesicles for the detachment of the 36S RNA from the template.

At the samesite of thisputative interaction,the presence of

protein

2C

(and

the other

2C-containing

viral P2proteins)

could consistentlybe demonstrated by

immunocytochemis-try (Fig. 5 and 9). Its presence in the replication complex even after DOC treatment indicates that it is not only

membrane associated but also boundto other components. This is in agreement with the findings that it is RNA associated and that it attaches viral RNAto the surface of the virus-induced vesicles

(6).

The 2C-RNA association is consistent with the

finding

that2C containsNTP-binding (8) aswellashelicase(12)motifs.Itmightwell be that 2C is also involved in the liberation of the RNA from the replication

complex

and

possibly

in preparatory steps forencapsidation

(16).

Whatever theexactfunctions of P2proteins2BC, 2B, and 2C

might

ultimately be, the action ofone or more of them wasfoundtobe essential for viralplus-strandRNA

synthe-sis

(6, 13, 15);

for

hepatitis

Avirus(9, 14), rhinovirus (17),

and

poliovirus (21) replication

in cell

culture;

and for expres-sion of

cytopathic

effects

(5).

Since these P2

protein-medi-ated processes are interconnected

(4),

one might speculate

that these

proteins

donotfunctionontheirownbut function

only in concert with each other and/or otherviral or even cellular proteins. This would also explain why, despite numerousattempts, e.g.,bymutationalanalysis,it is stillnot

possible

to attribute abiochemicallydefined roletoeach of these

proteins.

In our in vitro system, encapsidation of progeny RNA could not be observed. This is in contrastwith a recently

published

report

(18)

and

might

beduetoseveral

experimen-taldifferences. Our in vitro system isatranscriptionsystem, not a coupled transcription-translation system, and it was

kept going

foramuch shorter time. Animportant aspect of this

study

could be the

finding

that the 36S RNA is setfree from the replication complex immediately after its

comple-tion. This is in agreement with the

finding,

demonstratedby

in situ hybridization on infected cells, that a pool of free

genomic

RNAwaspresent in thevicinity ofthe replication

complex (27).

Inthe invitro system, thisRNAthenmightbe lost for encapsidation, because the RNA-synthesizing enti-ties are rather diluted in vitro compared with the in vivo situation.

The data presentedhereallowustopropose thefollowing model ofthe poliovirus replicationcomplex. The replication

complex

contains a compact membrane system which en-closes the RIwith its nascent plus-strand RNA. The repli-cation forks are in additiontightlyenclosed in protein. The compact membranes are involved in the liberation of the

completed 36S RNA from the RI and from the replication

complex.

The progeny strandsaresetfreeatdifferentpoints onthe

replication

complex's surface, i.e.,atthesites where the

replication

complex is attached to the virus-induced vesicles.

During

RNAsynthesis, theRI would thushave to move

(turn)

through, ortogether with, the replication

com-plex

with respect tothe virus-induced vesicles.

ACKNOWLEDGMENTS

Wethank Bea Herzog for excellent technical assistance. This work was supported bygrant 31-29910.90 from the Swiss National Science Foundation. M.T. wassupported byagrantfrom the Freiwillige Akademische Gesellschaft Basel, Basel, Switzer-land.

REFERENCES

1. Baltimore, D. 1968. Structureof thepoliovirusreplicative inter-mediate RNA. J. Mol. Biol. 32:359-368.

2. Bendayan, M., and E. Puvion. 1983. Ultrastructural detection of RNA:complementarityofhigh-resolution autoradiography and ofRNase-gold method. J. Ultrastruct. Res. 83:274-283. 3. Bienz, K., D. Egger, and L. Pasamontes. 1987. Association of

polioviral proteins of the P2 genomic region with the viral replication complex and virus-induced membrane synthesis as visualized by electron microscopic immunocytochemistry and autoradiography.Virology 160:220-226.

4. Bienz, K., D. Egger, Y. Rasser, and W. Bossart. 1980. Kinetics and location ofpoliovirus macromolecular synthesis in correla-tion to virus-inducedcytopathology. Virology 100:390-399. 5. Bienz, K.,D.Egger,Y.Rasser, andW.Bossart. 1983.

Intracel-lular distribution ofpoliovirus proteins and the induction of virus-specific cytoplasmic structures. Virology 131:39-48. 6. Bienz, K., D. Egger, M. Troxler, and L. Pasamontes. 1990.

Structuralorganization ofpoliovirus RNA replicationis medi-ated by viral proteins of the P2 genomic region. J. Virol. 64:1156-1163.

7. Dernick, R., and P. Bilello. 1974. Absence of glycoproteins in poliovirus particles.J. Gen. Virol. 25:125-132.

8. Dever, T. E.,M. J. Glynias, andW. C. Merrick. 1987. GTP-bindingdomain: three consensus sequence elements with dis-tinctspacing.Proc. Natl. Acad. Sci. USA 84:1814-1818. 9. Emerson, S. U., C. McRill, B. Rosenblum, S. Feinstone, and

R. H. Purcell. 1991. Mutations responsible for adaptation of hepatitis A virus to efficient growth in cell culture. J. Virol. 65:4882-4886.

10. Etchison, D.,and E.Ehrenfeld. 1981.Comparison of replication complexes synthesizing poliovirusRNA.Virology 111:33-46. 11. Girard, M. 1969. In vitro synthesis of poliovirus ribonucleic

acid: role of thereplicative intermediate. J. Virol. 3:376-384. 12. Gorbalenya,A.E.,E. V. Koonin, and Y. I. Wolf. 1990. Anew

superfamilyofputative NTP-bindingdomains encoded by ge-nomesof small DNAand RNA viruses. FEBS Lett. 262:145-148.

13. Johnson, K. L., and P. Sarnow. 1991. Three poliovirus 2B mutantsexhibitnoncomplementable defects in viral RNA am-plificationanddisplay dosage-dependentdominance over wild-typepoliovirus. J.Virol. 65:4341-4349.

14. Lemon, S. M.,P. C. Murphy, P. A. Shields, L.-H. Ping, S. M. Feinstone,T.Cromeans, and R. W. Jansen. 1991. Antigenic and geneticvariation incytopathic hepatitisAvirus variants arising during persistentinfection: evidence for genetic recombination. J.Virol. 65:2056-2065.

15. Li, J.-P., and D. Baltimore. 1988. Isolation of poliovirus 2C mutants defective in viral RNA synthesis. J. Virol. 62:4016-4021.

16. Li, J.-P.,and D. Baltimore. 1990. Anintragenic revertant of a poliovirus2Cmutanthasanuncoating defect. J. Virol. 64:1102-1107.

17. Lomax,N.B., and F.H. Yin.1989. Evidence for the role of the P2 protein of human rhinovirus in its host range change. J. Virol. 63:2396-2399.

18. Molla,A.,A. V.Paul,and E.Wimmer. 1991.Cell-free, de novo synthesisofpoliovirus. Science 254:1647-1651.

19. Pasamontes, L., D. Egger, and K. Bienz. 1986. Production of monoclonal and monospecific antibodies against non-capsid proteinsofpoliovirus. J. Gen. Virol.67:2415-2422.

20. Pfister,T.,L.Pasamontes, M.Troxler,D. Egger, and K. Bienz. Immunocytochemical localization of capsiol-related particles in subcellular fractions ofpoliovirus infected cells. Virology, in press.

21. Pincus, S. E.,D. C. Diamond, E. A. Emini, and E. Wimmer.

on November 9, 2019 by guest

http://jvi.asm.org/

1986. Guanidine-selected mutants of poliovirus: mapping of point mutationstopolypeptide 2C. J. Virol. 57:638-646. 22. Racaniello, V. R., and D. Baltimore. 1981. Cloned poliovirus

complementary DNA is infectious in mammalian cells. Science 214:916-919.

23. Sambrook,J., E. F. Fritsch, and T. Maniatis. 1989.Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

24. Takeda, N., R. J. Kuhn, C.-F. Yang, T. Takegami, and E. Wimmer.1986. Initiation of poliovirus plus-strand RNA synthe-sis in a membrane complex of infected HeLa cells. J. Virol.

60:43-53.

25. Takegami, T., B. L.Semler,C.W.Anderson, and E. Wimmer.

1983. Membranefractions active in poliovirus RNA replication containVPgprecursorpolypeptides. Virology 128:33-47.

26. Troxler, M., L. Pasamontes, D. Egger, and K. Bienz. 1990. In situhybridizationforlight and electron microscopy. A

compar-ison ofmethods for the localization of viral RNA using biotiny-lated DNA and RNA probes. J. Virol. Methods 30:1-14. 27. Troxler, M., T. Pfister, D. Egger,and K. Bienz. Submittedfor

publication.

28. Young, D. C., D. M. Tuschall, and J. B. Flanegan. 1985. Poliovirus RNA-dependent RNA polymerase and host cell protein synthesize products twice the size of poliovirionRNA in vitro. J. Virol. 54:256-264.

on November 9, 2019 by guest

http://jvi.asm.org/