Ends of the RNA within Sendai virus defective interfering nucleocapsids are not free.

(1)

0022-538X/78/0028-0584$02.00/0

Ends

of

the RNA

Within

Sendai Virus Defective

Interfering

Nucleocapsids

Are Not

Free

SUE LYNCH* ANDDANIEL KOLAKOFSKY

DepartmentofMicrobiology, University ofUtah MedicalCenter,Salt LakeCity,Utah 84132

Receivedfor publication 16 June 1978

Sendai virus defective interfering nucleocapsids, isolated from infected cell

cytoplasm by equilibrium banding in CsCl gradients, contain only the viral N

protein. Neither end of thegenomic RNA within thesenucleocapsidsisaccessible

toRNase digestion.

The genomeof the parinfluenza group of

vi-ruses is a single-stranded RNA surrounded

mainly by asingle protein, thenucleocapsid or

Nprotein, inahelicalmannersimilartothat of

tobaccomosaic virus(2,4,7).Thisstructure,the

viral nucleocapsid, is approximately 1

/m

long

when viewed in the electronmicroscope(11, 12)

andcontainsasingle-strandedRNAof5.0x106

daltons (15,000 nucleotides) or 5 ,im in length

(15, 16),complexed withapproximately3,000 N

proteins (60,000daltons),300Pproteins (70,000

daltons), and40Lproteins (-200daltons) (17).

The functions of theLandPproteinsare not as

yetknown, but it isthoughtthattheseproteins

are involved in the viral RNA polymerase or

transcriptaseactivity which ispresentinpurified

virionsand which has been localized in the

nu-cleocapsidcore(19, 21, 23). The structural

integ-rityof thenucleocapsidisprobably provided by

the interaction of the viral RNA and Nprotein,

sincenucleocapsidstreated withhigh

concentra-tionsof salt(-2M)lostmost,ifnotall,of their

LandPproteinsyetremain intactasjudged by

their appearanceunderan electronmicroscope

and theirbuoyant density in CsCl (19;

unpub-lisheddata).

The viral minus-strandgenomeisthoughtto

replicate via full-length complementary

plus-strand copy, the antigenome, since completely

double-stranded formsof the viral RNA which

are 5,umlongcanbeartificiallyconstructed from

nucleocapsid RNA (15). Suchcopies are found

in large amounts in the infected cell and to a

lesser extent in mature virion. These viralplus

strands orantigenomes,like the minusstrands,

are found only as nucleocapsids; no

genome-length viral RNA is ever found free in the

in-fected cell (5, 13, 22). The template for viral

replication is,therefore, thoughttobe a

nucleo-capsid,and theproductof thereplicativeevent

isthoughttobeacomplementarynucleocapsid.

Presumably,nucleocapsid assemblytakesplace

on the nascent RNA chain of the replicative

intermediate and continuesduring RNA chain

elongation sothatnofree genome-length RNA

iseverproduced. In the model ofgenome

repli-cation described above, the viral replicase

rec-ognizes a structure at the 3' end of both the

genome andantigenome nucleocapsids and

ini-tiates synthesis of the complementary RNA

strand at the 3' end of the nucleocapsid

tem-plate. The siterecognized by the viral replicase

toinitiatereplicationis thereforethoughttobe

specified at least in part by the 3' end of the

nucleocapsidtemplates.

To more clearlyunderstand thisrecognition

event, we have investigated the nature of the

ends of the viralnucleocapsids.Theexperiments

describedinthiscommunicationattempt to

dis-tinguish between two basicpossibilities: (i) the

RNA genome template is completely covered

with Nprotein,and the 3' end of the

nucleocap-sid RNA is not free, or (ii) a small number of

nucleotides at the 3' end of the nucleocapsids

are not complexed with N protein and remain

accessibleasfree RNAfor association with the

viral replicase. This latter possibility is

sug-gestedby therecent findingthat inUukeneimi

virus, another minus-strand virus of the

Bun-yaviridaegroup,the threenucleocapsidgenomes

appear circular when viewed in the electron

microscope (20), and their RNAs, when

ex-tracted,formcircularmolecules with short

pan-handles which appear to be held together by

complementary sequences at the ends of the

three genome RNAs (10). If the circular

Bun-yavirus nucleocapsids are held together by the

same complementary RNA sequences, this

would suggest that the ends of these genome

RNAs are notcompletelycovered with protein.

If asmallnumber of nucleotides are exposed

atthe ends of Sendai virusnucleocapsids, these

nucleotidesshould be available to RNase

diges-tion. However, since wild-type Sendai

nucleo-capsidscontain asingle RNA chain of

approxi-mately 15,000 N, the loss of a small number of

584

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

nucleotides would be difficult to detect. More suitable candidates on which to examine the

possibility that viral nucleocapsids contain free

RNA tailsareSendai virusdefective interfering

(DI)nucleocapsids. We have recently isolated a

1,200-nucleotide-long Sendai DI nucleocapsid

which, when disaggregated with sodium dodecyl

sulfate (SDS), liberates its RNA as

single-stranded circles held together by a 110-base

pair-long stem due to the base complementarity at

the ends of the DI RNA (14, 18). We have

therefore treated such nucleocapsids with

amounts ofpancreaticRNase sufficient to

com-pletely degrade free RNA and then examined

the remaining nucleocapsid RNA. This

com-munication reports that Sendai DI

nucleocap-sids so treated contain RNA indistinguishable

from untreated nucleocapsid RNA.

MATERIALS AND METHODS Preparation of Sendai virus DInucleocapsids. Confluent cultures of LLCMK2 cells were infected with a 1:10 dilution of fluid containing a mixture of both Sendai nondefective and H5 DIparticles (14). After45min, theinfecting mediumwasrecovered and

replacedwith3.5ml per100 mmof tissuecultureplate

ofminimalEaglemedium,containing 1% fetal bovine serum, 1

Ag

ofactinomycin D per ml, and 20uCi of

[3H]uridine perml (25 Ci/mmol). The infected cells

were harvested at 20 h postinfection, suspended in

0.15MNaCl,0.05MTris-chloride (pH 7.6), and 0.5% NP-40at aconcentration ofapproximately 25 x 106

cells perml,and thenrapidly disrupted inaVortex mixer. Nucleic andcellwallfragmentswerepelleted

bycentrifugation (5 minat3,000xg), and theresulting

supernatantwasmade5mM in EDTA.Samples (3ml

each) of this cytoplasmic extract were layered onto

CsCl-sucrose gradients (6.6 ml of20 to 40%[wt/wt]

preformed CsCl gradient and 2.3 ml of 5% sucrose

containing0.15MNaCl,0.02MTris-chloride, pH7.6,

and1mMEDTA)whichwerethencentrifugedfor16

hat30,000 rpminaSpinco SW40rotor.The visible

nucleocapsidband,whichwaslocatedatthecenterof

the preformed CsClgradient, was removedthrough

the sideof thecentrifugetube withasyringeinatotal volume ofapproximately0.3 mland storedat4°C.It has beenourexperience thatnucleocapsids storedin

30% CsClarestable foratleast1 month.For RNase digestionstudies,the CsClwasremovedby

chromat-ographing the nucleocapsids on Sephadex G-50 in

TNE(25 mMTris-chloride, pH7.4,50mMNaCl,and

1mMEDTA).

Polyacrylamide gel electrophoresis of RNAs.

RNAs were subjected to electrophoresis under

non-denaturing conditions on 8% acrylamide-0.4%

bis-acrylamide slab gels (1.5 by200by400mm)containing

40 mM Tris-chloride, 20mM sodium acetate, and 1

mM EDTA (pH 7.8). Ethanol precipitates of RNA

weredissolved in201dof one-fifththeconcentration of theelectrophoresisbuffer describedabove,

contain-ing 0.1% SDS; 10,l of 50% glycerol containing0.1%

bromophenol bluewasaddedtoeachsample, andthe

samplesweresubjectedtoelectrophresisfor 16 hat40 V.

RNAsweresubjectedtoelectrophoresisunder

den-aturing conditions on 1% agarose (SeaKem, Marine Collids, Inc.) slabsgels (3 by 200by400mm) contain-ingsodium borate bufferas describedbyBaileyand Davidson (1).Ethanol precipitates of RNAwere dis-solved in20,ulof5mM sodium borate(pH8.0),0.1% SDS, and 10 mM CH3HgOH; 10,l of 50% glycerol

containing 0.1%bromophenolbluewasaddedtoeach sample, and thesamplesweresubjectedto

electropho-resis for 16 h at 60 V. Before being processed for fluorography, the gelwasfixed in 10% acetic acid and

10mMcysteinetocomplextheCH3HgOh.

RESULTS

We havepreviously shown that viral

nucleo-capsids isolated from LLCMK2 cells infected

with a mixtureof wild-type and H5 DI particles

are composed almost entirely of a single DI

nucleocapsid species (14). This DI nucleocapsid has been shown to contain a 1,200-nucleotide-long RNA which sediments at 14S and

repre-sentsaunique segment of the viral genome (14,

18). When these DInucleocapsidsare

disaggre-gated with SDS and the RNA is isolated by

velocity sedimentation on sucrose gradients, the

RNA appears in the electron microscope as a

circular molecule heldtogether by a short (0.05

[m)

double-stranded stem. This

double-stranded RNA stem represents approximately

20% of the mass of the 14S DI genome, contains

both theoriginal 5' (pppA) and 3' (U-OH) ends

of the RNA chain and can be quantitatively

isolated fromthe circularsingle-stranded RNA

molecule by virtue of its resistance to RNase

digestion in 0.4 M NaCl (18). These Sendai DI

nucleocapsids arethereforeidealcandidates on

which to examine whether the 3' end of the

RNA is covered with Nprotein orremains

ex-posed.

Sendai H5 DI nucleocapsids were therefore

isolated frominfected cellcytoplasm by

equilib-rium banding in CsCl gradients as described

above, andthe CsCl wasremovedbygel

filtra-tion on Sephadex G-50. Triplicate samples of

the nucleocapsids in TNE containing 20,ug of

added rRNA per ml were digested with either

no (Fig. 1, panel a) 6

[Lg

(Fig. 1, panel b),or 18

,ug(Fig. 1,panel c) ofpancreatic RNaseperml,

andthe RNase was then inactivated with

die-thylpyrocarbonate (8). The nucleocapsids were

thendisaggregatedwithSDS, and the released

RNAs were sedimented on sucrose gradients.

Theresults (Fig. 1)showthat, althougha small

amountofnucleocapsidRNA isdegradedbythe

RNase digeston, the bulk of the

nucleocapsid-RNA continuestosedimentat14S. Asacontrol,

panel d of Fig. 1 shows the effect of 6 ,ug of

http://jvi.asm.org/

(3)

586 LYNCH AND KOLAKOFSKY

FRACTION NUMBER

FIG. 1. SDS-sucrose gradient sedimentation of

RNase-treatednucleocapsids. Triplicate0.2ml

sam-ples of[3H]uridine DI nucleocapsids (see text) in

TNEcontaining20jigofaddedribosomal RNAper mlwere incubated with0(a), 6(b), and18 (c) ,ugof

pancreatic RNase per mlfor 10min at25°C. The

RNase was inactivated by the addition of2 ,ul of

diethylpyrocarbonate, and the nucleocapsids were

then deproteinized by the addition of10 Il of10%o

SDS. Eachsamplewasthencentrifugedon5 to23%

sucrosegradients (17)for90minat58,000rpm in the Spinco SW60 rotor, thegradientswerefractionated,

and 10,ulofeachfraction was directlycounted by

liquid scintillation. The horizontal bars inpanelsa,

b, andcdenote thosefractions pooledasDI RNAs

(cf. Fig. 1, reference 18). The numbers above these horizontal bars representthepercentageof radioac-tive RNAin thepooledfractionswhich isresistantto

pancreatic RNase per ml on 30,g of["4C]rRNA

per ml under identical conditions. In addition,

RNasedigestion of the intact nucleocapsidsdid

notappreciably diminish the percentage of the

14S RNA which is further resistant to RNase

digestion due to its double-stranded character

(see numbers above horizontal bars in Fig. 1).

Since this RNase-resistant fraction represents

the complementary ends of the RNA chain,

these results indicate thatvery little, ifany, of

the ends of the DI RNA chain have been

re-moved by RNase digestionof the intact

nucleo-capsid.

A more precise way to examine whether any

RNA tails have been removed by the RNase

digestion is to examine the double-stranded

RNAstemsisolated from the14SDIRNA. We

havepreviously shown that H5DInucleocapsid

RNAyieldsa 110-basepair-longstemwhen the

single-stranded circular RNA is digested with

RNase in a high salt concentration (18). If

RNasedigestionof intact nucleocapsids has

re-moved a smaller number of nucleotides from

either end of the RNA,the length of the

double-stranded stem should be reduced accordingly.

Three-quarters of each pooled 14SRNA shown

in panels a, b, and c of Fig. 1 was therefore

treated with RNasein ahigh saltconcentration

(see legendto Fig. 2), and the remaining RNA

wassubjectedtoelectrophoresisas an 8%

poly-acrylamide slab gel.Asshown inFig. 2, panel a,

14SRNA derived fromnucleocapsidswhich had

been digested with RNase yielded a

double-stranded RNAstemwhose electrophoretic

mo-bility is indistinguishable from the

double-stranded RNA stem isolated from 14S RNA

derived from untreated nucleocapsids. Since a

difference ofeven 5 base pairs iseasily

distin-guishable on these gels (cf,,Fig. 2 of reference

18), these. results indicate that less than two

nucleotides could have been removed from

eitherend of the RNA chainbyRNasedigestion

of intactnucleocapsids.Todemonstrate that the

entire RNA chain in

nucleocapsids

is

protected

fromRNasedigestion,1/10 of each of the

pooled

14S RNAs shown inpanelsa,b, andcofFig. 1

was recovered by ethanol

precipitation,

dena-tured withmethylmercuryhydroxide,and

sub-jected toelectrophoresison anagarose slab

gel

containing methyl mercury hydroxide (1). As

pancreatic RNase digestion(30pg/ml)in0.3MNaCl

for30minat25°C. The dashed lineoverfractions1

to 5inpanelsa, b,andcrepresentsa10-fold

ampli-ficationoftheradioactivitypresent in thesefractions.

Panel d demonstrates theeffectof6jigof'4C-labeled LLCMK2 cell RNA per ml under identical conditions (opencircles);undigested'4C-labeled LLMCMK2 cell

RNA isdenotedbyclosed circles.

http://jvi.asm.org/

[image:3.499.63.253.55.495.2]

(4)

shown in panel b of Fig. 2, even though less

RNAwasrecoveredfromRNase-treated

nucleo-capsids, asingle RNA species of identical

elec-trophoreticmobilitywas detected in eachcase.

SincetheentireRNA chain in our DI

nucleo-capsidswas inaccessible toRNase digestion, it

wasof interesttodetermine which viral and/or

host proteins remained associated with the DI

nucleocapsids throughoutour purification

pro-cedure. Asample of the nucleocapsid

prepara-tion was thereforesubjected to electrophoresis

onan SDS-polyacrylamide slab gel along with

a sample of purified Sendai virus grown in

LLCMK2 cells. The gelwas stained with

Coo-massie brilliant blue and then scanned in an

Ortec densitometer. Panel b of Fig. 3

demon-stratesthatonly the viralNproteinandsomeof

the cellular contaminant actin (protein 7 [17])

remain associated with our DI nucleocapsids.

The association ofviral N proteinwith Sendai

genome RNA is therefore sufficient to protect

this RNA from RNasedigestion.

DISCUSSION

Theexperiments presented in thispaper

dem-onstrate that Sendai DI nucleocapsids isolated

from infected cell cytoplasm by equilibrium

banding in CsCl gradients (i) contain only the

viralNprotein, and (ii) neither end of the

gen-omic RNA within thesenucleocapsids is

acces-sibletoRNasedigestion.

The possibility that the ends of the RNA in

our DI nucleocapsid are not covered withNP

protein butarestillinaccessibletoRNase

diges-tion duetothecomplementaryends of the RNA

havingalready annealedtoformaduplex

struc-ture canbediscounted fortwo reasons: (i) when

viewed inanelectronmicroscope after negative

staining, thevast majority (99%) of the

nucleo-capsids appear aslinear and notcircular

struc-tures(unpublished data);and(ii) the pancreatic

RNasedigestionofintactnucleocapsidswas

car-riedoutunder NaCl concentrations (50mM)in

whichcompletely double-stranded MS2 RNA is

almosttotallydegraded (3). Since the viral

rep-licase isthought torecognize the 3' ends of the

genomeandantigenomenucleocapsid templates

during genome replication, these experiments

indicate that the viral replicase recognizes an

RNA-Nproteincomplexandnotfree RNA.

End-to-end aggregation of Sendai

nondefec-tive nucleocapsids in virus particles containing

multiple nucleocapsids, somereaching20 times

the unit length of 1 ,um, wasnoted as early as

1966 by Hosaka et al. (12). More recently,

Thorneand Dermott(24)havefoundrareforms

of what appeartobe circularmeaslesDI

nucleo-capsids. Inboth of these cases, the presence of

a b

1 2 3 --origin

2 3

-origin

_

-BPB

FIG. 2. (Panel a) Polyacrylamidegel electropho-resis ofdouble-stranded RNA stems from RNase-treated and unRNase-treatednucleocapsids. Of each of the

pooledDI RNAs denotedbythe horizontal bars in

panels a,b, andcofFig. 1, 75%weredigested with 20

pgofpancreaticRNase per ml in0.4M NaClfollowed

by Pronase digestion aspreviously described (18). The remaining RNA was twice precipitated with ethanol andsubjectedtoelectrophoresis on a poly-acrylamide slab gelasdescribed in the text. Thegel

wasprocessedfor fluorographyasdescribed by

Bon-nerandLaskey (6).(Panelb)Methyl mercury hydrox-ide-agarose gel electrophoresis ofRNase-treated and untreatednucleocapsidRNA. Ofeachofthepooled DI RNAs denotedby the horizontal bars in panels a, b, andcof Fig.1,10%/vwereprecipitatedwithethanol

andsubjectedtoelectrophoresison a1% agarose slab

gelcontaining10mMCH3HgOHasdescribed in the

text. Thegelwasthenprocessedforfluorographyas

previously described (6), except that methanol rather

than dimethylsulfoxide wasused todehydrate the

gel.

complementarityattheends of these

parainflu-enza nucleocapsid RNAs could explain these

end-to-end associations if the complementary

ends of the RNAswerefreetoanneal.However,

since the results presented in this paper

dem-onstrate that the complementary ends of the

RNA in Sendai DI nucleocapsidsare not

acces-ear'. 4%ft

http://jvi.asm.org/

[image:4.499.259.438.72.376.2]

(5)

588 LYNCH AND KOLAKOFSKY

DISTANCE MIGRATED

--FIG. 3. SDS-polyacrylamidegelelectrophoresis of nucleocapsid andmaturevirionproteins.Intracellular

viralnucleocapsids isolatedasdescribed in thetextwerepelleted bycentrifugation (60min,50,000rpmin the Spinco SW50 rotor),disruptedinsamplebuffer (10%/v glycine,5%/3-mercaptoethanol, 3%o SDS, and 0.06 MTris,

pH 6.8), andsubjectedtoelectrophoresison a 10%opolyacrylamideslabgel(panel a). PurifiedSendai virions grown in LLCMK2 cellsweresimilarly disrupted andsubjected to electrophoresison thegelasreference

markers (panel b). The gel was then stained with Coomassie brilliant blue and scanned in an Ortec densitometer.

sible to RNase digestion, it seems more likely

that the end-to-endaggregationofparainfluenza

nucleocapsids noted above is due tothe

inter-actionsof thenucleocapsid proteinsperse.This

wouldnotbeunusualforproteinswhichpossess

the abilityto self-assemble into helical

nucleo-capsids in a manner similarto that oftobacco

mosaic virus. End-to-endaggregationoftobacco

mosaic virus has also been shown to occur(9).

ACKNOWLEDGMENTS

This workwassupported bygrantPCM75-21043 from the National Science Foundation and Public Health Service grant AI-12921 from theNational Institute ofAllergyand Infectious Diseases.

LITERATURE CITED

1. Bailey,J.M., and N. Davidson. 1976.Methylmercury

as areversibly denaturing agentforagarosegel electro-phoresis.Anal. Biochem. 70:75-85.

2. Barry, R. D., and A. G. Buckrinskaya. 1968. The nucleic acidof Sendaivirusand ribonucleicacid syn-thesis incells infectedbySendai virus. J.Gen.Virol. 2:71-79.

3. Billeter,M.A., C. Weissman, andR.C.Wagner.1966. Replicationof viral ribonucleicacid:propertiesof dou-ble strandedRNA fromEscherichiacoliinfectedwith bacteriophage MS2. J. Mol.Biol. 17:145-173. 4. Blair, C. D., andW.S. Robinson.1968.Replicationof

Sendaiviruscomparisonof viral RNAandvirusspecific RNAsynthesiswithNewcastle diseasevirus.Virology 35:537-549.

5. Blair, C. D., and W. S. Robinson. 1970. Replication of Sendai virus. II. Steps in virus assembly. J. Virol. 5:639-650.

6. Bonner, W.M., and R. A.Laskey. 1974. Afilmdetection method for tritium-labelled proteins and nucleic acids inpolyacrylamide gels.Eur. J. Biochem. 46:83-88. 7. Compans, R. W., and P. Choppin. 1967. Isolation and

properties of the helical nucleocapsid of the parainflu-enza virus SV5. Proc. Natl. Acad. Sci. U.S.A. 57:949-956.

8. Fedorcsak, I.,and L.Ehrenberg. 1966.Effect of die-thylpyrocarbonate and methylmethioninesulfate on nu-cleic acids and nucleases. Acta Chem. Scand. 20:107-112.

9. Franki, R.I.B. 1962. Infectivity of tobacco mosaic virus from tobacco leavestreated with 2-thiouracil. Virology 17:9-21.

10.Hewlett, M. J., R. F. Petterson, andD. Baltimore. 1977. Circular forns of Uukuniemi virion RNA: an electronmicroscopic study. J. Virol. 21:1085-1093. 11. Hosaka,Y.1968.Isolation andstructure ofnucleocapsid

ofHVJ.Virology35:445-457.

12.Hosaka,Y., H.Kitano,andS.Ikeguchi.1966.Studies onpleomorphism of HVJ virions. Virology 29:205-221. 13.Kaverin,N. V. 1972.Delayin theincorporation of protein into virusnucleocapsid inNewcastledisease virus in-fectedcells. J.Gen.Virol 17:337-341.

14. Kolakofsky, D. 1976. Isolation and characterization of Sendai virusDI-RNA's.Cell 8:547-555.

15. Kolakofsky,D.,E.Boy de LaTour, andA.Bruschi. 1974. Self-annealing of Sendai virus RNA. J. Virol. 14:33-39.

16. Kolakofsky, D.,E.BoydeLaTour, and H. Delius, 1974. Molecular weight determination of Sendai and Newcastle diseasevirusRNA.J.Virol.13:261-269.

http://jvi.asm.org/

[image:5.499.114.404.68.305.2]

(6)

17. Lamb,R.A.,B. W. J.Mahy,and P. W.Choppin.1976. Thesynthesis of Sendai viruspolypeptidesin infected cells.Virology69:116-131.

18.Leppert, M.,L.Kort,and D.Kolakofsky.1977.Further characterization of Sendai virus DI-RNA's:amodelfor theirgeneration. Cell12:539-552.

19. Marx.P.A.,A.Portner, andD.W.Kingsbury. 1971. Sendai viriontranscriptase complex:polypeptide com-position and inhibition by virionenvelopeproteins.J. Virol. 13:107-112.

20. Pettersson, R.F.,andC.-H. vonBondorff. 1975. Ri-bonucleoproteins of Uukuniemi virus are circular. J. Virol. 15:386-392.

21. Robinson, W. S. 1971. Ribonucleic acid polymerase ac-tivity in Sendai virions and nucleocapsid. J. Virol. 8:81-86.

22. Robinson, W. S. 1971. Sendai virus RNA synthesis and nucleocapsid formation in presence of cycloheximide. Virology 44:494-502.

23. Stone, H. O., A. Portner, and D. W. Kingsbury. 1971. Ribonucleic acid transcriptases in Sendai virions and infected cells. J. Virol. 8:174-180.

24. Thorne, H. V., and E. Dermott. 1976. Circular and elongated linear forms of measles virus nucleocapsid. Nature (London) 264: 273-274.

http://jvi.asm.org/