Replication of vesicular stomatitis virus defective interfering particle RNA in vitro: transition from synthesis of defective interfering leader RNA to synthesis of full-length defective interfering RNA.

(1)

0022-538X/83/050513-10$02.00/0

Copyright C 1983,AmericanSociety forMicrobiology

Replication of Vesicular Stomatitis

Virus

Defective

Interfering

Particle RNA

In

Vitro: Transition

from

Synthesis

of

Defective

Interfering

Leader RNA

to

Synthesis

of

Full-Length

Defective

Interfering

RNA

GAIL W. WERTZ

DepartmentofBacteriology andImmunology, School ofMedicine, UniversityofNorth Carolina,Chapel

Hill, NorthCarolina 27514

Received 15November1982/Accepted25January1983

The replication of the RNA of vesicular stomatitis virus (VSV) defective interfering (DI) particles was established in a defined cell-free system. The transition from synthesis of only the DI-leader RNA to replication of the full-length DI RNA was effected in the systemby newly synthesized VSV proteins and occurred in the absence of VSV helper virus. Bothpositive- and

negative-polarityfull-lengthDI RNAweresynthesized. Furthermore,theproductsof RNA replication associated with newly synthesized viral proteins toform complexes thatwereindistinguishable from authentic DI particle nucleocapsidsonthebasis ofbuoyantdensity andresistancetoribonucleasedigestion. TheDI-leader RNA did not form ribonuclease-resistant structures. We conclude that this in vitro

systemsuccessfullyexecutesmanyof the reactions of VSV DI particle replication and assembly.

Oneof themajordifferences between thetwo

RNAsyntheticreactions, transcriptionand

rep-lication, that are carried out by the

negative-strand RNA virus, vesicular stomatitis virus (VSV), is thatreplication requiresviral protein synthesis, whereas transcription does not. The template for both RNAsyntheticreactions is the

negative-strandgenomicRNA(4 x 106 daltons)

in theform ofanucleocapsid structure, that is,

coated with the nucleocapsid protein, N, and

associatedwith thephosphoprotein, NS,and the large protein, L, which are components ofthe RNA polymerase. This structure is capable of

carryingouttranscriptionof thegenometoyield

leader RNAand the five VSVmRNAs(2, 10). Atpresent, we do not known precisely what

protein or proteins are required to effect and

maintain the transition from the synthesis of leader RNA and the discrete mRNAs (tran-scription) to the synthesis ofacomplete

read-through product to yield a full-genome-sized plus strand RNA, which issubsequentlyusedas

the template forsynthesis ofthe progeny

nega-tive-strandRNAgenomes(replication). Incells,

the full-length genomicRNAproducts of repli-cationarefoundonlyin the formof nucleocap-sids and are therefore resistant todigestion by

ribonucleases, whereas the mRNA products of

transcription are completely sensitive to

diges-tion by nucleases (30). Since the products of replicationareprotein-coatedRNAs, it has been

postulatedby numerousworkers thata

require-mentfor thenucleocapsid structural protein, N,

mayconstitute the need for continuous protein synthesis in negative-strand virusRNA replica-tion. It has been proposed that Nprotein may

play a specific role in catalyzing the transition from transcriptiontoreplication by bindingto a

site in leader RNAwhichmaybe thenucleation site for encapsidation (1, 14, 21). This event

would be dependent on the availability of N

protein and would determine the balance

be-tween replication and transcription. Ithasbeen proposed specifically that N functions to sup-press aterminationsignalatthe endof the leader

gene and that the binding of N within leader

simultaneouslystartsnucleocapsid assembly(5,

21).

Itis also possible that other newly synthesized

proteinsarerequiredtopromotereplication. For

example,theswitchtoreplicationmayrequirea newly synthesized L molecule, an L molecule

that has been modified by association with a

newly synthesized form ofthe phosphoprotein NS, which has been shown to exist in two

distinct phosphorylated forms (6, 15), orboth.

Additionally, itispossiblethat hostfactorsmay

playarole inreplication (24).

To investigate the requirement for protein synthesis in RNAreplication,adefined in vitro

system has been designed that supports both transcription and replication of the

negative-513

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

strand genomic RNA of the rhabdovirus, VSV (8). The system consists of three components: (i) purified VSV nucleocapsids as templates for RNA synthesis; (ii)an mRNA-dependent rabbit reticulocyte lysate to support protein synthesis; and(iii) purified VSV mRNAs to direct protein synthesis, if required. By using this combination of components, the level of viralprotein synthe-sis can be controlled as desiredby omission or addition of various amounts of viral mRNA. In

thissystem,replicationof thegenomicRNAis a

function of the level ofviral protein synthesis,

thereby allowing us to investigate the protein

requirements for replication.

Since VSVnucleocapsid templates carry out both RNA transcription and replication in this system, wechosetoextend our studies of

repli-cationby using a defective interfering (DI)

parti-cle asthe template forRNA synthesisto focus

solely on the process of RNA replication. The RNAof theDIparticleweselected (calledDI-T, VSI-DI 0.25; genome molecular weight, 0.9 x

106) contains only the 5' 25% of the standard virus genome;it completely lacks genetic

infor-mation forthe N, NS, M, and G proteins and contains approximately half of the information

for the L protein (19, 31). This DI particle, therefore, doesnotdirectmRNAsynthesis. The major RNA product made in vitro by the DI

particles generated from the 5' end of the genome is a 46-base RNA encoded by the ex-treme3' endofthe DIparticle RNA andcalled theDI-leader RNA(9, 28, 29). In cells coinfect-ed with DI particles and standard helper virus,

both the DI-leader andgenome-lengthDI RNA ofpositiveandnegative polarityaresynthesized (20, 25, 26). An important feature of the DI

particle RNA, for our purposes, is that the 3' terminus of thepositive strand (the template for

negative-strand replication)is identicaltothe3' terminus of the standard VSV positive strand

(13, 18). Inaddition, the 5' and 3' termini ofthe

DI RNAsarecomplementary. Therefore,the

3'-terminalsequencesof bothpositiveandnegative

RNA strands, which are the initiation sites for

RNA synthesis, are identical. Thus, this DI

particle constitutes an excellent template for investigating replication, since it is small,

con-taining only a quarterofthe genome, and does

notdirect mRNA synthesis, and yet it has the

correct sites for initiation ofreplicationand can

replicate efficiently in cells coinfected with in-fectiousVSV. Itisassumed that therequirement for helper virus is as a source ofmRNAs to

direct viral protein synthesis.

In this report we describe the replication in vitro of full-length DI particle RNA of both

positiveandnegative polarity.Thereplicationof

the DI RNA occurs in the absence of helper virus and only requires viralprotein synthesis.

Additionally, the full-length DI RNAs produced in this system are encapsidated with protein to form nucleocapsids that are indistinguishable from authentic DI nucleocapsids on the basis of buoyant density and resistance to ribonuclease digestion.

MATERIALS ANDMETHODS

Cell cultures and virus. Virus was propagated in

monolayer cultures ofBHK-21/13 cells as described

previously (33). The Indiana serotype(San Juan strain)

of VSVwasusedasstandardVSV. Stocks of the DI

particle, DI-T(VSI-DI0.25; 5' 25% of VSV genome

[31]), weregenerously provided by R. Lazzarini

(Na-tionalInstitutes ofHealth, Bethesda, Md.).

Preparation of nucleocapsids.Intracellular DI

parti-clenucleocapsidswereprepared byinfection of BHK

cellswith standard VSV at amultiplicityof infection of

1 andVSV DI particles (VSI-DI 0.25 was used in all

cases) at multiplicities indicated. DI nucleocapsids

were extracted from mixedly infectedcells by Dounce

homogenization, and DI particle nucleocapsids were

separated from standardvirusnucleocapsidsby

veloc-ity sedimentation in 15 to30%sucrosegradients (2 hat

38,500 rpm; Beckman SW40 rotor). The band of DI

nucleocapsids was collected and sedimented through

25% sucrose (2 h at 44,000 rpm; Beckman SW50

rotor), andthepelleted nucleocapsids were suspended

inHGD buffer(10%glycerol, 10 mM

N-2-hydroxyeth-ylpiperazine-N'-2-ethanesulfonic acid [HEPES], pH

7.6, 2 mMdithioerythritol).

Analysis of template RNA.The amount of unlabeled

nucleocapsid template RNA added to each reaction

wasquantitated by parallel preparationof

[3H]uridine-labeled templates with eachpreparation ofunlabeled

templates. The specific activity of labeled template

RNA was quantitated, and this figure was used to

calculate the amountoftemplate RNA in unlabeled

preparations made under identical conditions.

Similar-ly, the ratio ofpositive to negative strand full-length

DI RNAintemplatepreparationswasdetermined by

densitometricscanningoffluorogramsorbyexcision

ofappropriatebands fromgelsasdescribed below. In allexperiments where DIparticlenucleocapsid

tem-plateswereused,approximately1,000 pgofDI

parti-cletemplateRNAwasadded per25-,ulreaction.

In vitrosynthesis ofviral RNA and proteins. Viral

RNA and protein synthesis were carried out in the

presenceof amicrococcal nuclease-treated rabbit

re-ticulocyte lysate as described previously (8) except

thattheincubation timeswerevariedasindicated in

thefigure legends.

Analysisof RNAproducts.3H-labeled productswere

analyzed by electophoresis on 1.75% agarose-6 M

urea gels as described previously (32) or on 20%o

polyacrylamide slabgelsaccordingto the method of

Laemmli and Faure (16). DI-leaderRNA

(character-ized andkindly provided by L. A. Ball, University of

Wisconsin, Madison, Wis.)and DI particletemplate

RNA wereelectrophoresedasinternalmarkers inall

gels, and theirpositions areindicated. After

electro-phoresis, gels were fixed with 10% acetic acid and

subjectedtofluorography accordingtothemethodof

Laskey (17). Theresulting fluorogramswerescanned

toquantitate theintensityof bandsby usinganLKB

J. VIROL.

http://jvi.asm.org/

(3)

515

2202Ul trascan densitometer interfaced with anApple

II computer. Preflashed film was used to ensure a

linear response to low levels of radioactivity. Addi-tionally, a calibration curve of the linear response

range forfilm darkening was constructed and used to

ensurethatintensity scans were alwayscarriedoutin

thelinear range. To measure the amount of

radioactiv-ity in 3H-labeled RNA species, appropriate bands

were cut outof adried gel, using the fluorogram as a

guide.Thedried gelfragmentsweredissolved inNCS solubilizer (Amersham Corp.) at 60°C for 16 h before analysis by liquid scintillation spectroscopy.

Materials. [5,6_3H]UTP (specific activity, 35 Ci/

mmol)wasfrom ICN Pharmaceuticals.

RESULTS

Preparation of DIparticlenucleocapsids. VSV

DI paticles lack complete information for the structuralgenes, soit isnecessary topropagate

them by mixed infection of cells with standard VSV helper virus. The following method was usedto prepare DIparticle nucleocapsid prepa-rations that did not contain standard virus

nu-cleocapsidsto use astemplates forRNA

replica-tionreactions. Mixedinfections to propagate DI

particles were carried out at multiplicities that

yielded the greatest amount ofDI particle

nu-cleocapsids while giving minimum yields of stan-dard virus without totally inhibiting the infec-tion.Optimummultiplicitiesweredetermined by assaying the types of RNA synthesized inBHK

cells infected with standard virus at a

multiplic-ity of infection of 1 while the input of DI

particlesvaried. AsshowninFig. 1, decreasing dilutions of the DIparticle preparationgavehigh yields ofDIparticleRNAs, whereas synthesis of standard 40S virion RNA was greatly dimin-ished. These RNAs were analyzed on 1.75%

agarose-urea gels, which can resolve the posi-tive and negative strands of the DI particle genomicRNAs(11). The dilution of DI particles which wesubsequently employed inour

experi-ments was1:3,800,adilution which reduced the 12-hyieldof standard virus by98.6%, using the

Cooperand Bellettassay (7).

DIparticlenucleocapsidswereextractedfrom

mixedly infected cells andseparated from

stan-dard virus nucleocapsids by velocity gradient sedimentation. Thepurityof the RNAs isolated

by this procedure is shown in Fig. 2. Only the positive-andnegative-strandRNAsof DI

parti-cle nucleocapsids were detectable even after

overexposure offluorograms ofgels on which theseRNAs were analyzed. These results indi-cated that our preparations ofDI particle

nu-cleocapsids were essentially free of standard

VSV nucleocapsids and mRNA. Further

evi-dence that thepreparations were devoid of

de-tectable standard VSVnucleocapsidscamefrom

ananalysis oftheRNA products synthesized by

these preparationsin vitro (see below).

MIXED INFECTION

1

2

3

4 40S

RNA-L

RN

---

*

U

=

DI

RNA(|

-,_~~

"W

m--

NS-FIG. 1. Agarose-urea gel electrophoresis of RNA

from BHK cellsinfected with standard VSV and VSV

DI particles. BHK cellswere infected with standard

VSV at amultiplicity of infection of 1 and with VSV

DIparticlesatdilutions of: lane1,1:10,000; 2,1:7,500;

3, 1:5,000;and4,1:2,500. RNAlabeled with

[3H]uri-dinefrom thestartof infectionwasharvestedat12 h

postinfection fromcytoplasmic extractsandanalyzed byelectrophoresisin 1.75% agarosegels containing6 Murea.

Analysis of RNA products: effects of protein synthesis.RNAsynthesiswasdirected by the DI particle templates in the in vitro system, either in the absence orpresence ofconcomitant viral protein synthesis programmed by added VSV

mRNA. Theproducts of thereactionwere

ana-lyzed byelectrophoresison a1.75%agarosegel

containing 6 M urea (Fig. 3). This gel system

separatedthe DI particlegenome-size positive-andnegative-strandRNA aswellasretaining the 46-base DI-leader RNA and thereby allowed

simultaneous quantitation of both DI particle-specificproducts.

Themajor product synthesized by the VSVDI

particle nucleocapsids in the absence of viral protein synthesis was the 46-base DI-leader RNA (Fig. 3, lane 2). This finding confirmed previousreports that theDI-leader is themajor

product transcribed from the genomes of

puri-fied DI particles (9). In the presence of VSV

mRNA translation, two distinct changes in the

pattern ofRNAproductswereobserved:(i) full-length DI particle RNA of both positive and

http://jvi.asm.org/

[image:3.490.277.423.75.327.2]

(4)

Di

TEMPLATES

1

2

3 DI

RNA

H

particle genomic RNA as well.Analysisof pro-teinsynthesisinthis system(datanotshown)(8) has demonstrated that all five of the

VSV-specific proteins are synthesized. Synthesis of

theVSVproteinsis linear withtimefor up to 3 h and approximately 5 to 10 pmol ofprotein is madeduring180minina25-,ul reaction.Totest

directlywhether this transition in the pattern of RNA products was due to viral protein synthesis

or was a direct effect of the addition of viral

mRNA, protein synthesis was inhibited by the addition of cycloheximide orcycloheximideand anisomycin. Under these conditions, where pro-tein synthesis was inhibited by more than 99%, noDI full-length RNA was synthesized and the amountofDI-leader RNA made increased(Fig.

3, lane 4). These results showed that viral pro-tein synthesis isresponsible for enabling the DI

[image:4.490.75.220.62.310.2]

Di RNA

PRODUCTS

FIG. 2. Analysis of RNA fromDIparticle

nucleo-capsid preparations. DI particle nucleocapsids were

isolated from cytoplasmic extracts of BHK cells

in-fected with standard VSVat amultiplicityofinfection

of 1 and DI particles at adilution of1:3,800bytwo

cycles ofvelocity sedimentation.RNA wasextracted

from nucleocapsids and analyzed as described in the

legendtoFig. 1. Lane 1, Marker RNA from DIparticle

virions; 2, RNAfrom DI particle nucleocapsids

(fluor-ogramexposed 3days);3,same aslane 2 except that

thefluorogram was exposedfor7days.

negative polarity appeared (Fig. 3, lane 3), whichcomigrated with markertemplateDI

par-ticle RNA (Fig. 3, lane 5); and (ii) therewas a

marked reduction in the amount of DI leader that was produced. The appearanceof the full-length RNAswasdependentonthepresencein the reaction mixture of all four ribonucleoside

triphosphates. This findingindicated that these products were synthesized de novo and that

they were not generated by terminal incorpo-rationorexchange of labeled nucleotide.

At high concentrations of DI particle nucleo-capsids, the synthesis of full-length DI RNA

becameless efficientand, in someexperiments,

was notcompletelydependent on the addition of viral mRNA. The reasons for this areunknown,

but onepossibilityis that thehigh concentration ofviralproteinscontributedbythenucleocapsid preparation was able to support a limited

amountofRNAreplication.

Thedatapresentedabovesuggested that pro-tein synthesis directed by VSV mRNAs was

responsible forthe transition from synthesizing only leader RNA to synthesizing full-length DI

1 2 34

5 DI RNA

= -- (+)

DI-Leader

additions:

mRNA

cyclo

0++

0./0+

FIG. 3. Agarose-urea gelelectrophoresis ofRNA

products synthesized byDI particlenucleocapsidsin

vitro.RNAproducts were labeled with[3H]UTP in the

cell-free systemprogrammed withthefollowing

com-ponents:lane 1,noDItemplates; 2, DItemplates, no

VSV mRNA; 3, DI templates, VSV mRNA; 4, DI

templates, VSV mRNA, and cycloheximide (50

,g/ml); 5, markerDItemplateRNA. Incubationwas

for 180 minat4°C. RNAwasextracted andanalyzed

asdescribed inthelegendtoFig.1.

J. VIROL.

http://jvi.asm.org/

[image:4.490.277.423.284.575.2]

(5)

REPLICATION OF

particlenucleocapsidstocarryoutthesynthesis of full-length RNA copies.

Quantitation of RNA products. The reduction in DI-leader synthesis under conditions where thereplication offull-length RNAwasoccurring

promptedus toquantitate theamounts of these

two products. The relative amounts of full-length DI RNA and DI-leader RNAwere

quanti-tated by scanning laser densitometry of

fluoro-grams of gels of the separated products. The

datarepresented in Table 1 show that for each 1 mol offull-length DI product (both positive and negative strand) produced in the presence of

protein synthesis, approximately 100 mol of DI-leaderwas synthesized. In the absence of

pro-tein synthesis, threefold more DI-leader was

synthesized than in the presence of protein synthesis. If cycloheximidewasaddedtoinhibit protein synthesis, more leader was made,

al-though less than was seen in the absence of added mRNA and inhibitor. These resultsarein agreementwiththeobservationthatthe amount

of negative-strand leader in infected cells is

increased in thepresence ofcycloheximide (5).

The ratio ofpositive to negative strand for full-sizedDIRNAproductwascomparedtothis

ratio forthe input template nucleocapsidRNAs.

Averaged data from five separate experiments showed that after 90minofsynthesis, the

prod-uct DI full-length RNA was composed of

ap-proximately 42% negative-strand RNA and 58% positive-strand RNA. Input template

prepara-tions contained anaverage of 57%negative and 43% positive strands. These data showed that synthesis of progeny full-sized RNA in this

systemclosely reflected the ratio ofinput

tem-plate RNAsatearly times in the reaction.Thisis theresultonewouldpredict if the full-length DI

productwascopied from the input template. At

later times in the reaction (3 h), the ratio of positive- to negative-strand products became

approximately equal. This finding may suggest

thatRNAssynthesizedatearly timescan serve as new templates for replication atlater times.

Additionally, the absoluteamountofinput

tem-plate RNAwascalculated andcomparedtothe

amountof DIfull-lengthRNAproduct by

exci-sion ofappropriate bands from gels. Approxi-mately1,000pgoftemplateRNAwasaddedper

25-IlI reaction andapproximately 400to700pg

ofDIgenome-lengthRNAwasproduced.

Kineticsof RNA synthesis. To investigate the

decreaseinDI-leadersynthesis with theonsetof RNA replication, we analyzed the kinetics of appearance ofbothDI-leader andfull-length DI RNA. For these experiments, the synthesis of DI-leaderRNAwasanalyzed byelectrophoresis ofsamples on20%oacrylamide gels followed by

[image:5.490.254.448.76.174.2]

densitometric scanning of fluorograms of dried gels(Fig. 4). Initially,DI-leaderwasproducedat

TABLE 1. Relative molar amountsof DI particle

RNAproducts

Molar amt Reaction DIgenome- DI-leader

sizedRNAb RNA

DInucleocapsids alone 0 328

DInucleocapsids + VSV mRNA 1 93

DInucleocapsids + VSV mRNA 0 114

+ cycloheximide _I

a Relative molar amounts were calculated by

densi-tometric scanningof fluorograms of[3H]UTP-labeled

products separated by gel electrophoresis. Uridine

contentsof6.5% for DI-leader (28, 29) and31%for DI

particle genome RNA (27) were used in making calcu-lations.

bBoth positive- and negative-strand genome-sized

RNAs wereincludedin this calculation.

thesame ratewhether mRNAhadbeen addedto programviral protein synthesisor not.

Howev-er,after 20to40min,while DI-leader continued

toaccumulateat alinearrate in the absence of protein synthesis, therewas amarked decrease in its rate of accumulation in the presence of

ongoing protein synthesis. Indeed, there was

little additional DI-leader RNA accumulation after 40 min in reactions with ongoing protein synthesis.

Acorrespondingkineticanalysisoffull-length DI RNA synthesis showed that full-size RNA

was notdetectable until approximately60 to 90 min after the startof the reaction (Fig. 5). The accumulation of this productwasnotlinear with time, and the major accumulation of full-length DIRNAoccurredabruptly2 to3 h after thestart

of the reaction.

Association of RNA products with protein. StudiesofRNAreplicationin infected cells have shown that replication is dependenton protein

synthesis and that both positive- and

negative-strandgenome-sizedRNAsarefoundonly in the form of nucleocapsids, never as naked RNA. Having demonstrated that full-sized DI particle

RNAcould be synthesized in vitro and that its

productionwasdependentonprotein synthesis,

we nextexamined theability of the RNA

prod-ucts to associate with newly synthesized pro-teins. Theassociation was assayed in two ways.

First,thebehavior ofthe RNAproducts in CsCl density gradientswas analyzed. The

3H-labeled

DIfull-length product banded in CsClgradients

atthesamepositionasmarker DInucleocapsids (datanot shown), indicating thatthe RNA was notnaked but wasassociated withprotein and, also, that the RNA/protein ratio was

approxi-mately the same as thatof authentic

nucleocap-sids. Next, the susceptibilityof the RNA

prod-uct to digestion with ribonuclease was

VOL.46,1983

http://jvi.asm.org/

(6)

DI-LEADER

a b c d e f

_-.origin

-._

_-w

am-DI-Leader

+ m0Ni

After 10 minat23°Cethylene

glycol-bis(Q-amin-oethylether)-N,N-tetraacetic acid was added to inhibit further nuclease activity. The reactions were terminated, and the RNA was extracted and analyzed inthe same manner as the undi-gested samples. Nuclease digestion of mRNA transcriptionproducts synthesized in the invitro

system by VSV virion nucleocapsids was per-formed under identical conditions to test

wheth-erthemicrococcal nuclease was active in these conditions. The results of this experiment are presented in Fig. 6.Only the full-length DI RNA products were resistant to nuclease digestion (Fig. 6, lane 6). The DI-leader RNA was digest-edby micrococcal nuclease even when made in the presence of ongoing protein synthesis. As

expected, the VSV mRNAs synthesized by

nucleocapsids from infectiousvirionswerealso

SYNTHESIS OF

DI

RNAs

A..

1

minutes 6 9 2 8

at 30 0 0 0 0 M

==

82+DI

[image:6.490.51.239.70.428.2]

minutes at 30(

FIG. 4. Kinetics of accumulation of DI-leader

RNA. (A) DI-leader RNAsynthesiswascompared in

cell-free reactions programmedwith DI templates to

which VSV mRNAwas(b,d, andf)or wasnot(a,c,

and e)added. RNAproducts synthesizedin the

pres-ence of [3H]UTP added atthe start of the reaction

wereanalyzed at 20(a,b), 40(c,d),or80(e,f)minby

electrophoresis on 20% acrylamide gels. 3H-labeled product comigrated exactly with marker DI-leader

(position indicated). (B) Fluorogramsof the driedgels

were scanned with a laser densitometer, and the

relative amounts ofproductsynthesized duringeach

timeinterval wereplotted.

examined. RNA products were synthesized in

vitro in the presence orabsence of viralprotein synthesisasdescribedforFig.3.After90minof

incubation, the reaction mixtures weredivided

in half. Half of each reaction was immediately terminated by the addition of sodium dodecyl sulfate. The RNA was extracted and analyzed by gel electrophoresis. The other halfofeach

reactionmixture wasdigestedwithmicrococcal

nuclease in the presence of calcium chloride.

15-E

C3

10-0

E

'Z

5-a

6(0 20 120 lSC

minutes at 30

FIG. 5. Kinetics of appearance offull-lengtb DI

particle RNAs. (A)[3H]UTP-labeledproducts

synthe-sized in the in vitro system programmed with DI

templates and added VSVmRNAwereanalyzedat60,

90, 120, and180minafter thestartof thereaction by

electrophoresis as described in the legendtoFig.1. A

small amount of positive- and negative-strand

full-length product was detectableat60miinontheoriginal

fluorogram,but was not of great enoughquantitytobe

discernedonsubsequentphotographic exposure. Lane

Mshows marker DItemplate RNAs. (B) Fluorograms

of the dried gel were scanned as described in the

legendtoFig.4, and thekinetics of appearanceof

full-lengthproduct (bothpositive and negative strand) are shown.

A.

C c

0

c

C)

a)

B.

/~~~~~~~~~~~~~~~----J. VIROL.

http://jvi.asm.org/

[image:6.490.256.445.283.529.2]

(7)

519

Nuclease Sensitivity of Product RNA

Control

DI B

123 4

- -L

Nuclease Treated

DI B

56 7 8

indicate that they contain N, NS, and L, the three viral proteins found inauthentic VSV and DI particle nucleocapsids (22).

DISCUSSION

.r

m

,i",p

:>

uaciitlonsu;

mRNA 0++

[image:7.490.49.239.74.363.2]

cyclo 0 0+

FIG. 6. Nuclease

sizedby DI particle

RNAproducts synth

sids in thein vitros)

inthepresenceofno

5), added VSV mRI

mRNAplus cyclohe)

[3H]UTP-labeled pr(

VSVnucleocapsidsv

4and8)were analy2

after lanes 5 throug

nuclease,(10Fg/ml)

phoresisasdescribec

completely digeste 8).

These results de length DI positive

are synthesizedin

thesis-dependentri

teins made are a

newlysynthesized

are stable in 7 M

density gradientsN

cleocapsids, andti

digestion with nu4

characteristic ofni

such structures ar tem. Furthermore, found in these n

The results presented above demonstrate that it ispossible to carry out the replication of full-length VSV DI particle RNAs ofboth positive

*-G and

negative

polarity

in vitro. The ratios of

@

ZN

positiveto

negative

strandprogenyRNA

closely

reflected the ratio of input nucleocapsid

tem-NS

plate RNA at earlytimes in thereaction. At later M times (3 h), approximately equal amounts of

positive and negative strand RNAs were pro-duced. The DI particle nucleocapsid templates

were able to make the transition from synthesiz-ingonly the DI-leader RNA in vitro to synthesiz-ing full-length, DI genome-sized RNA by viral protein synthesis. This transition was blocked byinhibitorsof protein synthesis. These results

show that DI RNA replication is regulated by

viral protein

synthesis

in this system, as itis in

Aeader. ~~~~infected cells.

The events involved in replication of DI RNA can beconsidered in more detail. In the absence

o 0+ + 0 of

protein

synthesis,

the DI-leader is the

only

0 00+ 0 product synthesized by the DI nucleocapsid template. A strong termination signal is present

sensitivity of products synthe- at the endoftheleadergene. Theoretically,any

nucleocapsids. [3H]UTP-labeled read through or elongation ofproduct beyond

iesizedby DIparticlenucleocap- theDI-leaderboundarycanbe considered

repli-ystemduring a 90-minincubation cation. The successful elongation of the product

)addedVSVmRNA (lanes1 and to yield a full genome-sized RNA, however, may NA (lanes 2 and 6), added VSV be aprocess involving different protein require-ximide (50,ug/ml)(lanes 3 and 7). ments than the initial read-through event. For

oducts synthesized by standard example, the transition from transcription to

with noadded VSV mRNA (lanes

eplcathe

mayrequire tic

orichiometo

zed before(lanes 1 through 4) or

replication

mayrequire catalytic or stoichiomet-,h 8) digestion with micrococcal ric (or both) amounts of a certain protein(s) or a for 10 min at23°C by gel electro- short-lived intermediate (see above), whereas dinthelegend to Fig. 1. the efficient elongation ofproduct to yield full-length progeny may require only a constant

supply of the nucleocapsid protein N. We have notassayed directlyfor theformer eventinthis

dby thenuclease(Fig. 6, lane system at this time. Rather, we chose to assay only for the complete read-through product, a

mmonstrate not only that full- full-lengthDI molecule. Therefore, based on the >- and negative-strand RNAs studies presented here, we can say that for

thissystem in a protein syn- efficient synthesis of full-length product to oc-eaction, but also that the pro- cur, a certain concentration of viral protein must

tble to associate with these be available. The findings reported here are RNAs to formstructuresthat consistent with the requirement for continuous I CsCl, that coband in CsCl proteinsynthesistomaintaingenome RNA syn-withauthentic DIparticle nu- thesis in VSV-infectedcells (23,33).

hatarecompletely resistantto A striking aspect of the

quantitation

ofRNA

clease. These properties are productswas the finding that as protein synthe-ucleocapsidsand suggestthat siswasinitiatedin the system and the synthesis eformed in this in vitro sys- offull-length DI RNA occurred, the amount of analyses of the viralproteins DI-leaderRNA decreased threefold. When the ewly synthesized structures kinetics of appearance of the DI-leader RNA

VOL.46, 1983

_>of:s:..

http://jvi.asm.org/

(8)

were compared in the presence and absence of viral protein synthesis, it was observed that the rateofaccumulation was similar for the first 20 minof the reaction. After that time, however, while DI-leader continued to accumulate at a linear rate in the absence ofprotein synthesis, there was a marked decrease in itsaccumulation in the presence ofprotein synthesis.The appear-ance of full-length DI RNAcould not be detect-ed untilapproximately 60to90min inreactions in whichprotein synthesiswas initiatedattime 0. These datashowed that the sharpdecrease in therate of accumulation of DI-leader RNA was

followed by the appearance of full-sized DI RNA. These two events,however, did not occur untilapproximately 40 min after the start of the reaction.

Inconsidering these results, it is important to note that viralprotein synthesis was being pro-grammed from the start of the reaction and,

therefore, that nopreexisting poolof free virus proteins existed. One explanation for the results of these kinetic studies may beasfollows. After protein synthesis is initiated in the system and proteinsbegin to accumulate, anecessary

con-centration ofprotein is achieved such that the transition can be made fromtermination atthe endofthe DI-leadertoproductionoffull-length product. As this occurs, one might predict a

decrease inthe appearanceofdiscreteDI-leader

RNA as this molecule is elongated to produce

the full genome-sized product. However, this

explanationcanaccountforonlyasmall part of the decrease in accumulation of discrete DI-leader (Table 1). We do not know what other factors may be involved. It ispossiblethat the rateof initiation in the system slows downasthe result ofproduction of some new viralprotein,

orperhapstheinvolvementofpolymerase mole-cules in elongation of the DI product restricts the amount ofpolymerase available for initia-tion. Anotherpossibilityis that there isapause

atthejunctionbetween DI-leader and therestof the DIgenome,which may result inaslowingof

polymerase movement and new initiation

events. Iverson and Rose (12)haveshown that thereisadistinctpause ofapproximately 5min in the rate of VSV RNA transcription ateach intercistronicjunction. Taken together,the sim-plestexplanation of these results and onethatis consistentwith our knowledgeofreplication in theinfected cell is thatthetransitionto replica-tion offull-length RNAin this system requires

notsimplyongoing protein synthesisbut rather theavailability ofacertain concentration of viral protein before it ispossibletodetect the synthe-sis of the full-sized RNA.

Theworkdescribed abovealsodemonstrated that the full-length DI positive- and

negative-strand RNAs, made in a protein

synthesis-de-pendentreaction, could associate with the pro-teins synthesized concomitantly in the system. The RNA productsbanded in CsCl at the same buoyant density as marker DI nucleocapsids, indicating that the RNA was associated with proteinsand that the RNA/protein ratio was the sameas that ofauthentic nucleocapsids. More-over,both the full-length positive- and negative-strand RNAproducts were completely resistant todigestion with ribonuclease. Our data showed that all of thereplicated product was protected from nuclease digestion. Since the electropho-reticmigration of the RNAs after nuclease diges-tion was the same as that before, these data suggestthat, within the limits of this assay, the full-length RNAs were fully protected. These findings show that the newly synthesized RNA and proteins associate to form structures that resembleauthentic nucleocapsids.

Quantitation offull-length DI product RNA demonstrated that approximately 1.4 to 2.1 pmol ofnucleotide was synthesized in a 25-,lI reac-tion. In this same reaction, approximately 4.8

pmol of viral protein was made, ofwhich 1.7 pmol was N protein. Assuming that the size of the DI genome RNA is approximately 3,000 bases, these data show that there are roughly 2,500 to3,600 N moleculesavailableper mol of DI product RNA. Bishop and Roy (3) have calculated that thereareapproximately2,000 N molecules per standard VSV genome. The DI genome is approximately one-quarter this size and, based on the above figure, should

require

approximately500Nmolecules to coat each full-size DI genome. Thus, the replication system reported here produces sufficient N protein to coat the progeny DI RNAs. This finding is consistent with thefact that all the DIfull-length

RNAis nucleaseresistant,and itshows that this system can produce viral proteins and RNA in quantities appropriate to successfully carry out RNAreplication andnucleocapsidassembly.

The DI-leader RNA, in contrast to the full-lengthDI RNA, however, was not resistant to

digestionwith ribonuclease. Amodel has been put forward (5, 21) which proposes that the

availability of N protein controls the balance between replication and transcription by the

binding ofN protein to a site in leader RNA, whichthereby attenuates the termination at the

end of leader and allows read through while

simultaneously creating a nucleation site for

encapsidation;failuretobindNresultsin termi-nation and release ofleader is adiscrete

mole-cule. Weobservedthatthe accumulationof free DI-leaderRNA wasdecreased inthepresence of protein synthesis and full-length RNA

replica-tion. Moreover, we have shown that all full-length DI RNA is encapsidated with protein, whereas free leader is not. These findings are

J. VIROL.

http://jvi.asm.org/

(9)

compatible with, but donot prove,theidea that Nprotein binding to aleader RNA canresultin elongation to yieldgenome-length RNA, where-as failure of N to interact with leader results in termination and release offree leader. In

con-trast to our finding that free leader was not

encapsidated in our system,Blumberg and Kola-kofsky (4) reported that approximately three-quarters of the total intracellular leader RNA

was encapsidated. In relation to the model

pro-posed, theirfindingsuggests thatassociation of

N with leader may not always result in read throughor thatencapsidation offreeleadermay

occur after release from the template in the infected cell, or both. We suggest thatin our in

vitro system, RNAsynthesis occurs under con-ditions where proteinconcentration may not be

as great as in the infected cell. Under these circumstances, it may be that free leader is revealed as a less efficient competitor for N

protein than a nascent leader molecule still at-tached to its template.

In conclusion, an in vitro system has been

establishedwhich supportsthe synthesis of full-length VSV DI particle RNA and carries out its encapsidation to form nucleocapsids. In this system, the transition fromsynthesis of only the DI-leader RNA, a transcription event, to the replication offull-length DI RNA is made as a

function of viralprotein synthesis, whichcanbe controlled by omission or addition of viral mRNA. Thus, individual VSV mRNAs can be added alone or in combination to program pro-tein synthesis, and the effects of theseproteins on the transition to RNA replication can be evaluated. This approach providesamethod for assaying the involvement of individual proteins in the process of RNAreplication.

ACKNOWLEDGMENTS

I thank Nancy Davis for helpful advice anddiscussions, John Glass for artwork, and Steve Loechel for excellent technical assistance.

This research was supported by Public Health Service grants AI12464 and AI15134 from the National Institute of Allergy andInfectious Diseases and grant CA19014 from the National CancerInstitute.

LITERATURECITED

1. Ball, L. A.1980.Geneorder,p.61-73. In D. H. L.Bishop

(ed), Rhabdoviruses.CRC Press, Inc.,Cleveland,Ohio. 2. Banerjee, A., G.Abraham,and R.Colonno.1977.

Vesicu-lar stomatitis virus: mode oftranscription. J. Gen. Virol. 34:1-8.

3.Bishop, D. H. L., and P. Roy. 1972. Dissociation of vesicular stomatitis virus and relation of the virion pro-teins to the viral transcriptase. J. Virol. 10:234-243. 4. Blumberg, B. M., and D. Kolakofsky.1981. Intracellular

vesicular stomatitis virus leader RNAs are found in nu-cleocapsid structures. J. Virol. 40:568-576.

5. Blumberg, B., M. Leppert, and D. Kolakofsky. 1981. Interaction of VSV leader RNA and nucleocapsid protein may control VSV genome replication. Cell 23:837-845. 6. Clinton, G. M., B. W. Burge, and A. S. Huang. 1978.

Effects ofphosphorylationandpHontheassociation of

NS protein with vesicular stomatitis cores. J. Virol.

27:340-346.

7. Cooper, P.,andA.Belett.1959. A transmissable

interfer-ingcomponent of vesicularstomatitisviruspreparations. J.Gen.Microbiol. 21:485-497.

8. Davis, N. L., and G. W. Wertz. 1982. Synthesis of

vesicularstomatitis virusnegative-strandRNAinvitro: dependenceonviralproteinsynthesis. J. Virol. 41:821-832.

9. Emerson,S.U.,P. M.Dierks,andJ.T.Parsons.1977. In vitrosynthesisofauniqueRNAspecies byaTparticleof vesicular stomatitis virus. J. Virol. 23:708-716. 10. Emerson, S. U. and Y.-H. Yu. 1975. Both NS and L

proteins are required for in vitro RNA synthesis by vesicularstomatitisvirus. J. Virol. 15:1348-1356. 11. Huang, A., D. Rao, and G. Lanman. 1980. Defective

interfering particlesofvesicularstomatitis virus:structure

function relationships. Ann. N.Y.Acad. Sci.

354:238-250.

12. Iverson,L.,andJ.Rose.1981. Localized attenuation and

discontinuoussynthesisduringvesicular stomatitis virus

transcription.Cell 23:477-484.

13. Keene, J.,M.Schubert,R.Lazzarini,andM.Rosenberg. 1978. Nucleotide sequencehomologyatthe 3'terminiof

RNA from vesicular stomatitis virus and its defective

interfering particles. Proc. Natl. Acad. Sci. U.S.A. 75:3225-3229.

14. Kingsbury,D.1974.The molecularbiologyof paramyxo-viruses. Med.Microbiol.Immunol. 16:73-83.

15. Kingsford, L.,andS. U.Emerson.1980.Transcriptional

activities of differentphosphorylated speciesof NS

pro-tein purifiedfromvesicular stomatitis virions and

cyto-plasm ofinfected cells. J. Virol. 33:1097-1105. 16. Laemmli, U.,and M. Faure.1973.Maturation of the head

ofbacteriophageT4. I. DNApackagingevents.J. Mol.

Biol. 80:575-599.

17. Laskey, R. 1980. The use of intensifying screens or

organicscintillators forvisualizingradioactive molecules

resolved by gel electrophoresis. Methods Enzymol.

65:363-371.

18. Lazzarini, R., J. Keene, and M. Schubert. 1981. The

originsofdefectiveinterferingparticlesof the

negative-strand RNAviruses. Cell26:145-154.

19. Leamnson, R.,andM. E.Reichmann.1974. The RNA of

defectivevesicular stomatitis virusparticlesin relationto viral cistrons. J. Mol. Biol. 85:551-568.

20. Leppert,M.,andD.Kolakofsky.1980.Effect of defective

interfering particles on plus- and minus-strand leader

RNAs invesicular stomatitis virus-infected cells. J. Virol. 35:704-709.

21. Leppert,M.,L.Rittenhouse,J.Perrault,D.Summers,and

D.Kolakofsky.1979.Plus and minus strand leader RNAs

innegativestrand virus-infected cells. Cell 18:735-747. 22. Patton, J. T., N. L.Davis,andG.W.Wertz. 1983.

Cell-freesynthesisandassemblyof vesicular stomatitis virus

nucleocapsids.J.Virol.45:155-164.

23. Perlman,S.M.,and A.S.Huang.1973. RNAsynthesisof

VSVV.Interaction betweentranscriptionandreplication.

J.Virol. 12:1395-1400.

24. Prlngle, C. 1978. The tdCE and hrCEphenotypes: host

range mutants of vesicular stomatitis virus in which

polymerasefunctionisaffected. Cell15:597-606.

25. Rao, D.,andA.Huang.1979.Synthesisofasmall RNA in cells coinfected by standard and defective interfering

particlesof VSV. Proc. Natl. Acad. Sci. U.S.A.

76:3742-3745.

26. Rao, D., and A. S. Huang. 1980. RNA synthesis by vesicular stomatitis virusX.Transcriptionandreplication

bydefectiveinterfering particles.J. Virol. 36:756-765. 27. Repik, P.,and D. H. L.Bishop.1973.Determination of the

molecularweightof animal RNA viral genomesby

nucle-asedigestions.J.Virol. 12:969-983.

28. Schubert, M., J. Keene, R. Lazzarini,-and S. Emerson. 1978. The complete sequence ofaunique RNA species

http://jvi.asm.org/

(10)

synthesized byaDIparticle of VSV. Cell 15:103-112. 29. Senler, B., J. Perrault, J. Abelson,andJ. Holland. 1978.

Sequence of a RNA templated by the 3' OH RNA terminus of defective interfering particles of vesicular stomatitis virus. Proc. Natl. Acad. Sci. U.S.A. 75:3923-3931.

30. Soria, M., S. Little, and A. Huang. 1974. Characterization ofvesicular stomatitis virus nucleocapsids. I.

Comple-mentary40S RNA molecules in nucleocapsids. Virology

61:270-280.

31. Stammnger,G., and R.Lazzarinl. 1979. Analysis of the

RNA of defective VSV particles. Celi 3:85-93. 32. Wertz, G. W., and N. L. Davis. 1979. RNase IIIcleaves

vesicular stomatitis genome-length RNAs, but fails to cleaveviral mRNA's. J. Virol. 30:108-115.

33. Wertz,G. W., and M. Levine. 1973. RNA synthesis by vesicular stomatitis virus and a small plaque mutant: effectsof cycloheximide. J. Virol. 12:253-264.

http://jvi.asm.org/