Mechanism of attenuation of a chimeric influenza A/B transfectant virus.

(1)

Vol. 66, No. 8 JOURNALOFVIROLOGY,Aug. 1992, p. 4679-4685

0022-538X/92/084679-07$02.00/0

Mechanism of

Attenuation of

a

Chimeric Influenza

A/B Transfectant Virus

GUANGXIANGLUO, MICHAEL BERGMANN, ADOLFO GARCIA-SASTRE,t ANDPETERPALESE*

Departmentof Microbiology, MountSinaiSchoolof Medicine,

1 Gustave L. LevyPlace, New

York,

NewYork10029-6574

Received 27 March 1992/Accepted 19April 1992

The ribonucleoprotein transfection system for influenza virus allowedus to construct an influenza A virus

containingachimeric neuraminidase (NA)genein which thenoncodingsequenceis derived from theNS gene of influenza B virus(T. Muster, E. K.Subbarao,M. Enami, B. P. Murphy, and P. Palese, Proc. Natl. Acad. Sci. USA88:5177-5181, 1991). This transfectant virus is attenuated in mice and growstolower titers in tissue culture thanwild-type virus.Since such a virus has characteristics desirable foraliveattenuated vaccinestrain, attemptsweremadetocharacterize this virus at the molecularlevel. Our analysis suggests that the attenuation of the virus is duetochanges in the cis signal sequences, which resulted in areduction oftranscription and replication of the chimeric NA gene. The majorfindingconcerns a sixfoldreductioninNA-specific viral RNA

inthe virion,causing a reduction in the ratio of infectious particles to physical particles compared with the ratio inwild-type virus. Although the NA-specific mRNA level is also reduced in transfectantvirus-infected cells, it doesnotappear tocontribute to theattenuationcharacteristics of thevirus. The levels of the other RNAs and theirexpression appeartobeunchanged for the transfectant virus. Itis suggested thatdownregulation of the synthesis of one viral RNA segment leadstothegeneration ofdefective viruses during each replication cycle. We believethatthis represents a general principle for attenuation which may be appliedtoother segmented virusescontaining either single-stranded or double-stranded RNA.

Influenza, anacuteinfectious respiratory disease, remains

a constant worldwide threat to human health and, in the absence ofeffective antiviral agents, can best be prevented

or limited byvaccination (1, 13). The influenza virus vac-cines currently in use in the United States are derived from inactivated influenza virus, and they provide only limited protection in vaccinees (2, 13). An alternative

ap-proachwould involve the development of safe live attenu-atedvaccines; these vaccines should be longer-lasting and

afford greater protection (1). An ideal way to generate suchanattenuatedvaccinewould beto usegenetic engineer-ing techniques. However, until recently, influenza

virus-likeall othernegative-strand RNAviruses-wasnot

amena-ble to site-specific genetic manipulation. The situation has

changed for influenzavirus,becausesyntheticRNAs

recon-stituted in vitro withpurifiedviralpolymerase complexcan now be rescued intoinfectious influenza virus particles (5,

10, 12).

Applying thisnovel reverse genetic approach, we gener-ated a transfectant influenza virus which possesses a chi-meric neuraminidase (NA) gene containing the coding

se-quenceof theNAgeneofinfluenzaA/WSN/33virus andthe

noncoding sequence from the NS segment of influenza B/Leevirus. The viruswas showntobeattenuated inmice

and to induce a protective immune response against

chal-lenge virus. These features suggested that the virus has many properties desirable for an attenuated live influenza virus vaccine (14). In order to understand the molecular characteristics responsible for the attenuation of this

trans-fectant virus, we carried out a series of experiments to

*Correspondingauthor.

tOnleaveofabsencefromtheDepartmentofBiochemistryand MolecularBiology, FacultyofBiology, University ofSalamanca, Salamanca, Spain.

analyze the virusatthe molecular level. The data presented inthisreport suggestthatattenuation results from the effect of the altered cis elementsonthereplication of the chimeric NA gene.Alow level ofthe NA gene in the virus preparation results in a higher proportion of defective particles than is foundin awild-typeviruspreparation. Inaddition, the data support a random mechanism for the packaging ofvirion RNAs(vRNAs)into influenza virus particles.

MATERIALSANDMETHODS

Viruses and cells. Thestocks ofinfluenzaA/WSN/33virus and the NA/B-NS transfectant virus were prepared from

purified plaques by growing them in Madin-Darby bovine

kidney (MDBK) cells in reinforced minimal essential

me-diumcontaining 2 ,ug oftrypsin per ml. MDBKcells were used for the plaquing of viruses and the study of

virus-specific RNAsynthesis.

Plasmids. In order to generate riboprobes, several

plas-midswereconstructed. pSP64-NS DNA contains the entire NS segment ofA/WSN/33virusinserted into the SalI site of the pSP64vector(Promega, Madison, Wis.) in the orienta-tionbywhichpositive-senseNS RNAcanbe madebyusing

SP6 RNApolymerase. To create plasmid pIBI30-NS, the entire NS DNAfragment derived from pSP64-NS was in-serted between theHindIII and EcoRI sites of the pIBI30 vector(IBI,NewHaven,Conn.)in the orientationbywhich

negative-senseNSRNA canbegenerated byusingT7RNA polymerase. In order to obtain plasmid AT3NAv, which produces the negative-sense NA-specific RNA probe, the

DNAin thepT3NAvplasmid(5)was shortenedby deleting

nucleotides 1 to 1096 of the NA vRNA. In

addition,

we

created a plasmid, designated IBI30-NA, by

inserting

the entireNAfragmentderived frompT3NAv

(5)

into theXbaI 4679

on November 9, 2019 by guest

http://jvi.asm.org/

(2)

4680 LUO ET AL.

and EcoRI sites of the IBI30 vector. This plasmid can produce positive-sense NA-specificRNAbyT7 RNA poly-merasetranscription.

Virus purification and RNA extraction. Influenza A/WSN/33 virus and NA/B-NS transfectant virus were growninMDBKcells and thenpurified by30 to60%sucrose gradient centrifugation as described previously (5). Virus

purifiedfrom four175-cm2flasks of MDBKcellswas

resus-pended in0.3 mlof TMK buffer (10mM Tris [pH 7.5], 1.5 mMMgCl2, 10 mMKCl) anddisrupted byincubationwith 9 pul of 10% sodium dodecyl sulfate (SDS) and 7.5 pul of

proteinase K (10 mg/ml) at 56°C for 10 min, followed by

addition of 35 ,ulof SLN buffer(5% SDS, 1.4 M LiCl, 100 mMsodiumacetate [pH 7.0]). Virion RNAswereextracted with phenol-chloroform and collectedbyethanol precipita-tion. For isolating viral RNAs from infected cells, MDBK cellswere infected with either influenzaA/WSN/33virusor NA/B-NStransfectant virusatamultiplicityofinfection of 1 and harvested at the indicated time points. Cells were

washed twice with ice-cold phosphate-buffered saline and

lysed with 4 M guanidinium isothiocyanate (Sigma). Total RNAwas then purified by equilibrium centrifugationin 5.7 Mcesium chloride (Fisher) (9).

Determination ofthe ratio of infectious andphysical parti-cles. In order to determine the total number of physical particles ofinfluenza A/WSN/33virus andNA/B-NS trans-fectant virus, the virus preparations were mixed with an equal volume of a suspension of carboxylate polystyrene beads (0.1 pum in diameter) at aconcentration of4.5 x 109 particles per ml (Polyscience, Inc., Warrington, Pa.) and then stained with phosphotungstic acid. Theratio of virus andpolystyrenebeadswasdeterminedby countingthe two differentparticles under the electronmicroscope. For mea-suringthenumber of infectiousparticlesin thepreparations, the virusstockswereseriallydiluted andplaquedinMDBK cells.

RNA electrophoresis. RNAs extracted from influenza A/WSN/33virus andNA/B-NStransfectantviruswere

elec-trophoresed on a 3% polyacrylamide gel containing 7.7 M ureaat150 V for 2 h.The RNAsegmentswerevisualizedby silverstaining as described previously (5).

RNaseprotection assay. The RNaseprotection assaywas used forquantitationof virion RNA in viralparticlesaswell as for the measurement of viral mRNAs and cRNAs in infected MDBK cells. The NS segmentwas chosen as an internal control. FormeasuringvirionRNAs, positive-sense NS- andNA-specificRNAprobesweregenerated byrun-off

transcription with phage SP6 or T7 RNA polymerase and plasmid pSP64-NSDNA linearizedbyNcoI(nucleotide [nt] 265) or plasmid IBI30-NA digested with FokI (nt 240), respectively. The NS-specific RNAprobe spans theregion between nt 1 and265 of the NScRNA,and theNA-specific RNAprobe covers nt 1 to 240 ofthe NAcRNA.

Todetermine the amounts of mRNAsandcRNAs, nega-tive-sense NS- and NA-specific RNA probes were tran-scribed from DdeI (nt 465)-digested IBI30-NS DNA and PvuII (nt 306 in pUC19)-cut AT3NA DNA, respectively, with phage T7 or T3 RNA polymerase, respectively. The negative-sense NS-specific RNAprobe is470nt long, con-tainingnt465to890 of theNSvRNAandavectorsequence

of45 nt. The negative-sense NA-specificRNAprobeis403 nt long, coveringtheregionbetweennt1096 and1409 of the NAvRNAand containing a vectorsequence of 90nt.

The RNA probes were labeled with [a-32P]UTP (800 Ci/mmol; DuPont, NEN, Boston, Mass.). Ingeneral,50ng

of virion RNA extracted from purified virus was

hybrid-ized to 5 x 104 cpm each of the positive-sense NS- and NA-specific probes, or 5 p.g of total RNA isolated from virus-infected cells was hybridized to 5 x 104 cpm each of the negative-sense NS- and NA-specific probes. After 12 h ofincubation at 45°C, the hybridization mixture was digested with RNases A and

Tl,

following the manufactur-er'sinstruction (Ambion Inc., Austin, Tex.), and the result-ing productswereanalyzed on a 6% acrylamide denaturing gel (9).

Primer extension. The genomic RNAs (vRNAs) of the NA and NS segments of influenza A/WSN/33 virus and of

NA/B-NS transfectant virus were quantitated by primer extension (11). The primers for detecting NS- and NA-specific vRNA are 21 nt long and are complementary to

negative-sense vRNA. The NS primer, 5'-GGGAACAATT

AGGTCAGAAGT-3', spansthe region between nt 695 and 714 of the NS cRNA. The NAprimer, 5'-GTGGCAATAAC TAATCGGTCA-3',coversnt1151to1171 of the NA cRNA. Both the NS and NA primers were 5'-end labeled by incubation with[_y32P]ATP (3,000 Ci/mmol; Du Pont, NEN) and T4 DNA kinase(Biolabs, Beverly, Mass.). Then, 100 ng of RNA extracted from virus or 5 ,ug of total RNA isolated from infected MDBK cells was reverse transcribed by RNase H-minus murineleukemia virusreversetranscriptase

(BRL, Gaithersburg, Md.) in the presence of 3 x 105 cpm each of the5'-end-labeled NSand NA primers. After incu-bationat37°Cfor2h,thereactionwas stopped by addition of EDTAto 10 mM, and the reaction mixwassubjected to

phenol-chloroformextraction and alkalitreatment(11). The

productswereanalyzedon a6%polyacrylamide gel

contain-ing 7 M urea. The amount of product was measured by

directly counting the radioactivity of the gel piece corre-spondingtoeach band onthe film.

Neuraminidase assay and Western immunoblot analysis. For the assay of neuraminidase activity of the influenza

A/WSN/33

andNA/B-NStransfectantviruses,

2'-(4-methyl-umbelliferyl)-at-D-N-acetylneuraminicacid(Sigma)wasused

asthe substrate. Thereaction mixture consisted of 25 ,ul of 2 mMsubstrate,25 ,ulofvirus,and50 ,u of 0.2 Mphosphate

buffer(pH 6.0) containing2 mM CaCl2.Afterincubation at

37°C

for 10min,thereactionwasstopped byaddition of 2 ml of 0.5 M glycine-NaOH buffer (pH 10.6), and then the neuraminidase activity was determined by measuring the

fluorescencewithexcitation at365 nm and emission at450 nm, with methylumbelliferone as a standard. The protein concentration of the viruseswasmeasured with the Bio-Rad

proteinassay kit. For the Western analysisofthe NA and HA proteins of the WSN/33 and NA/B-NS viruses, viral

proteins were electrophoresed on a 10% polyacrylamide

Laemmligel (7) andsubsequentlytransferredto a nitrocel-lulose membrane (Schleicher & Schuell, Keene, N.H.). A monoclonal antibody directed against carbohydrates was

usedtodetect theNAandHAglycoproteinsof the viruses. TheWesternblotwasdevelopedwitha ratantibody against mouse kappa chains, which was labeled with 125I and was

generously provided byThomas Moran.

Analysis ofprotein synthesisin infectedcells. MDBKcells

(35-mm dish) were infected with either WSN/33 virus or NA/B-NStransfectant virus at amultiplicity of infection of

approximately 3. This multiplicity was used because the

NA/B-NStransfectantvirus didnotgrowtohigher titer. At indicated times, the proteins were labeled in cysteine-free medium with

[35S]cysteine

(1,027 Ci/mmol;DuPont, NEN) at 100 ,uCi/ml of medium for 30 min. The cellswere then washed twice with ice-cold phosphate-buffered saline and lysed in 150

pul

oflysisbuffer containing 1% Nonidet P-40,

J. VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

(3)

ATTENUATION OF CHIMERIC INFLUENZA A/B VIRUS 4681 NA of Influenza A WSN/33 Virus

5AGUAGAAACAAG AQUUUGAACAAACUA

3- UCGCUUU C GUCCUCAAAUJIAC

NA of NA/B-NSTransfectant Virus

5AGUAGUAACAAGAGGAuuuuuAUUUUACAUUCA

3UUCCUUCGUCUCCUA

AAUAAAUCAGUGACCGUUUGCCUUUCUAC

FIG. 1. Noncodingsequencesof the NAsegmentsofinfluenza

AIWSN/33virus and the NA/B-NStransfectant virus. The 5'- and 3-terminal sequences are drawn in a panhandle structure, which consists of twobase-pairedstemsand onemismatchedinternalloop in themiddle. The noncoding nucleotides of thechimericNA gene of the NA/B-NS virus arederived from the NS gene ofinfluenza B/Lee virus. The large letters indicate nucleotidesin the 5'- and

3'-terminal regionswhicharedifferentfor thetwo NA genes. The opentrianglemarks the altered Kozaksequencein theNA geneof theNA/B-NS virus.

150 mM NaCl, 50 mM Tris-HCl (pH 8.0), and 1 mM phenylmethylsulfonyl fluoride. About 1/20 of the sample was loadedonto a 10% polyacrylamide Laemmli gel (7).

RESULTS

Ratio of infectious to physical particles.The genome of the transfectant NA/B-NS virus differs from that ofwild-type influenza A/WSN/33 virus onlyin the noncoding region of the NA gene (Fig. 1). It is thus likely that the altered

biological properties of the transfectant virusarethe result of alteredcis signals located in thenoncoding regionof the chimeric NA gene (9, 15, 18). Specifically, it was noted earlier that the transfectant virus grewto lower titers than wild-type virus in MDBK cells, MDCK cells, and mice. In addition, the low-multiplicity growth curves in tissue

cul-ture weresignificantly delayed relativetothoseofwild-type

virus (14). We then proceeded to examine whether the

transfectant virus had a temperature-sensitive phenotype,

which could explain the altered growth characteristics.

However, the pattern of the growth curve at 30, 33, 37,

and 38°C for NA/B-NS virus was not different from that ofwild-type virus at the corresponding temperatures. We

next asked whether defective particles were present in theNA/B-NSvirus

preparation.

Theviruswascharacterized by counting the physical particles under the electron

microscope and comparing this number with the PFU of the preparation. Interestingly, the NA/B-NS transfectant virus showed a similar number of physical particles as

the wild-type virus but consistently lower PFU titers (14) (Table 1). Thus, theNA/B-NS virus grown in MDBK cells shows at least a 5- to 10-fold-lower infectious

particle-to-physical particle ratio than is seen with the WSN/33 wild-typevirus.

[image:3.612.60.300.682.727.2]

Characterization of theRNAof the NA/B-NS virus. Since

TABLE 1. Ratioof infectiousparticlestophysical particlesof influenzaA/WSN/33virus andNA/B-NStransfectant virus Virus Physical_{(pp) (no./ml)}particles Infectious particles Ratio, pp/ip

(ip) (no./ml)

WSN/33 1.8 x 109 1.2x 108 15

NA/B-NS 1.2 x 109 1.0 x 107 120

the NA/B-NSvirus contains many defective particles, the viral RNA was examined for the presence of defective RNAs. Following extraction from purified virus, the ge-nomic RNA was separated on a 3% polyacrylamide gel

containing 7.7 M urea. As shown in Fig. 2A, the chimeric NA RNA of the NA/B-NS transfectant virus is almost invisibleon the gel, whereas theother sevensegments are

present in approximately equimolar concentrations. When theamountof RNAonthe gelwasincreased,the chimeric

NARNAwasshown tomigrateatthesamepositionasthe control RNAin lane3(datanot shown). However,

electro-phoresis andsilverstainingdidnotpermitthequantitationof

thechimericNA RNApackagedinvirions.For this purpose, the RNase protection assay and primer extension experi-ments wereperformed.

Positive-sense NS-specific and NA-specific probes were hybridizedinthesamereactiontopurifiedvirion RNA from

eitherWSN/33 virus(Fig. 2B,lane3)orNA/B-NS transfec-tantvirus(Fig. 2B, lane4) and then digestedwith RNases A and T1. The resulting products were analyzed on a 6% polyacrylamide gel

containing

7 Murea

(9),

andtheamounts

of vRNAwerecalculatedby countingtheradioactivityof the

gel slices

corresponding

tothe bandson the film. As shown

in Fig. 2B, the probe protected bythe chimericNA RNA

migratesfaster than thatof thewild-typegenebecausethe 5

noncoding sequences are different in the two NA genes.

Compared with the NA segment of WSN/33

virus,

the

amount of chimeric NA RNA in the transfectant virus is about sixtimeslower, with the NSgene usedasaninternal

control.The

primer

extension

experiment,

asshown in

Fig.

2C, showsasimilarreductionin theamountofchimericNA RNApackagedinvirions relativetotheNS gene,suggesting

a

specific

lower

representation

of the chimericNA gene in

thetransfectant virus preparation.

Characterization of theproteinof theNA/B-NSvirus. The

nextquestionwaswhether theNA/B-NStransfectantvirus also contained less neuraminidase

protein.

We carried out neuraminidase assays and a Western

analysis

of the NA

protein in virus particles. For the enzymatic assayof the

neuraminidase, the totalconcentration of viral

proteins

was

first determined by the

protein

assay, and then the same amount of

purified

viruswas used. The

NA/B-NS

transfec-tantvirus showedonlyatwofold-lower NA

activity

thanthe WSN/33 wild-typevirus. A similar

finding

was obtainedby

Western analysis, which showed 1.9-fold less NA

protein

but the same amount of HA

protein

in virions of the

NA/B-NS transfectant virus

compared

with those of

WSN/33virus(see Fig.

5A).

Virus-specificRNAsynthesisininfectedcells.Basedonthe vRNAanalysisof theNA/B-NStransfectant

virus,

there is little doubt that lesschimeric NA RNA is contained in the viralparticles. Thelower level of the chimeric NARNAin the virions could be caused either

by

a

change

in the

packaging signal, leading

to less efficient

packaging

of the

NARNA, orby a changein the

synthesis

of

genomic

NA RNA,whichcouldbe the result ofinefficient

recognition

of the influenza B virus-specificpromoter by the influenza A

viral polymerase. To

pinpoint

the exact

mechanism,

we

examined the vRNA

synthesis

of the chimeric NA gene in infected MDBK cells.

Considering

that vRNA is

mainly

synthesized

in the late

phase

of infection

(19),

weusedRNA

extracted from cellsat7,

8.5,

and 10 h

postinfection.

Total RNA(5

,ug)

extracted fromvirus-infected cellswasused for theprimerextension

analysis.

The data in

Fig.

3show that vRNAsynthesis ofthe chimeric NAgenewas

remarkably

decreased relative tothat of control virus.

Compared

with

VOL.66,1992

on November 9, 2019 by guest

http://jvi.asm.org/

(4)

4682 LUO ET AL.

A.

WSN/33 NA/B-NS

eI

Control

F....

i-NA/B-NS C.

310 281 27 1 234 194

B.

(0c Un o c. Z .0a) *

*_

Z

_n

M

L-0

a. O ulz

NS

NA d Z NS

0

-NA

+4NAIB-NS

2 3 4

X n

s z)

)-.

4-NA

[image:4.612.115.480.78.303.2]

.--NS

FIG. 2. Characterization of theRNA oftheNA/B-NS virus. (A)RNAelectrophoresis. The RNAs extracted from purifiedviruseswere analyzedon a3%polyacrylamidegelcontaining7.7 M urea andvisualized by silver staining. Lane 1, RNA of influenza A WSN/33 virus; lane 2, RNAofNA/B-NStransfectantvirus; lane 3, RNAobtained by run-off transcription fromplasmidpT7NA/B-NS, which produces the chimericNARNA.(B) Analysis ofNARNA invirions by the RNase protection assay. RNA (50ng)extractedfrompurified virus was used in thehybridization reactionwith positive-sense NS-andNA-specific riboprobes asdescribed in Materials and Methods. The protected probeswereelectrophoresed on a6% acrylamide gel containing 7 M urea. Lane 1, riboprobes without RNase A orT1digestion; lane 2, riboprobes following RNase A andT1digestion; lane 3, riboprobes protected by the RNA of influenza A/WSN/33 virus; lane 4, riboprobes protected by the RNA ofNA/B-NS transfectantvirus. (C) Quantitation of NA-specific RNA in virus by primer extension. The vRNA extracted from either WSN/33 virus (lane 2) or NA/B-NS transfectant virus (lane 3) was reverse transcribed by RNase H-minus reverse transcriptasewithNS andNAsegment-specificprimersasdescribed in Materials and Methods. The products for the NS RNAs are 195 nt long,andthose forthe NA RNAs areapproximately 260ntlong.The products wereanalyzedona 6%polyacrylamide gel containing 7 M urea. Size markers are shown onthe left(in nucleotides).

the synthesis of the NA RNA of WSN/33 virus, synthesis from thechimericNA geneis reduced by a factor of 9, 8, and 8at7,8.5, and 10 hpostinfection,respectively. However, no

reduction of vRNAsynthesis was observed with respect to

w)

WSN/33 NA/B-NS 1 7 8.510 1 7 8.510

Vicnr

3 Z

Virion _p.i.(hr)

310

281 4-NA

271

234

23

4

194w O*, .*

~ ~ ~ 4

FIG. 3. Analysis of NA-specific vRNA synthesis in infectedcells by primerextension. RNA isolated from infected cellswasreverse

transcribed with reverse transcriptase and NA and NS vRNA-specific primers, asdescribed in Materials and Methods. vRNAs extracted from purified viruswere used as a control (right). The

resulting products were displayed on a6% acrylamide denaturing gel. Thereversetranscripts of the NS RNAsare 195 ntlong, and those of theNA-specificsegmentsareapproximately260ntlong,as

indicatedbyarrows atthe right. The numbersonthetopindicate times postinfection. Virion represents the RNA obtained from purified virus. Sizemarkersareshown in nucleotides.

the NS segment of the NA/B-NS transfectant virus. This result suggeststhat the reduction of thechimeric NA RNA in transfectant virus is similar to the reduction in

synthe-sis observed in infected cells. Thus, the lower

representa-tion of the chimeric NA segment in virus particles is most

likelynotdueto adefect in the packaging of the chimeric NA RNA.

In ordertodetermine the level of mRNAsynthesis of the chimeric NAgene, MDBK cellswere infected with either WSN/33 virus or NA/B-NS transfectant virus, and total RNAwasthen isolated from virus-infected cellsatdifferent timespostinfection. Subsequently, the level ofvirus-specific

mRNAs was quantitated by the RNase protection assay.

Again, the NSsegment wasusedasacontrol.From the time course, it isapparent thatthe level ofchimeric NA mRNA is markedly reduced relativetothat of thewild-type virus and that it only increases slightly with time (Fig. 4A). At 5 h postinfection,thelevel ofchimericNAmRNA wasfivefold

less than that of WSN/33 virus (Fig. 4B), whereas the NS-specific mRNA synthesis of the NA/B-NS transfectant

virus was similar to that ofWSN/33 virusat the indicated times (Fig. 4C). It should be noted that the NS-specific mRNA synthesis appears earlierand is more efficient than NA-specificmRNAsynthesisinbothWSN/33andNA/B-NS

transfectant virus-infected cells (probes ofsimilar activity

were used in the experiment). This finding is in good

agreement with previous reports (4) and suggests that the mRNA synthesis of differentinfluenzavirusRNA segments

isdifferentially regulated. The conditions of theassaydidnot J.VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.77.277.500.624.2]

(5)

ATFIENUATION OF CHIMERIC INFLUENZA A/B VIRUS 4683

A

NS Probe - 470

(-)Sense Sense

N.S...-RNA.... NS mRNA

5

'-Product

AAA(A)n 3'

+ 409

nt*

NA Probe

t=03

4 nt

(-)Sense 3 '

NA mRNA 5

'-Product

IAAA(A)n 3'

14-298

nt-W'

a) 0

QD

-o0

-CL C)

NSo:

NA-o

v

.Vl

WSN/33

Virus

1 3 4 5 6 7

3

lo45

eOQe

NA/B-NS

Virus

1 3 4 5 6 7

Hours(p.i.)

-NS

4-NA

B 4000 --- WSN/33

* NA/B-NS

C

a

._

0

EL

cJ

2000

1000

4 6

Hours (p.i.)

-0--O WSN/33

-a- NA/B-NS

FIG. 4. Timecourseof mRNAsynthesisin MDBKcells.(A) Quantitation ofNA-andNS-specificmRNAsatdifferent timespostinfection (p.i.) bytheRNaseprotectionassay.Thediagramonthetopschematically illustrates the procedure.Both theNSand NAprobesarenegative senseand containsequencescorrespondingtothe 3'-terminalside of thecRNA, flankedbyvectorsequencesatthe 3'terminus,asindicated by theopenrectangles.The sizes oftheprobesareshownonthetop, and the sizes ofthe resulting productsareindicatedatthe bottom of thediagram.The timepointsareindicatedonthe top. Thepositionsofprobes and productsonthegelareindicatedbyarrowsatthe left and right, respectively. (B)TimecourseofNA-specificmRNAsynthesis.Theamountof the NA mRNAwasmeasuredby directly countingthe radioactivity (countsperminute)in thecorrespondingband excised from thegelshown inpanelA. (C) ComparisonofNS-specificmRNA synthesisoftransfectant virus andA/WSN/33virus in infected cells. Theamountof mRNA foreach timepointwasdeterminedasdescribed forpanelB.

permitustoquantitatethe level of cRNA of the chimeric NA

genebecausethe cRNAsynthesiswasfoundtobeonly3to 5% of that of mRNAsynthesis (lOa).

Reduction of NAprotein synthesisinvivo. The NAprotein was found to be reduced by a factor of 2 in virions of the NA/B-NS transfectant virus, whereas thesynthesis of chi-meric NA mRNAwasreducedbymuchmore.Thequestion

thenarosewhether the NAproteinencodedbythe chimeric NA mRNA inside cellsparallelstheamountof itsmRNA,or whether thereisaselective incorporation of the

neuramini-dase into the viral envelope. In orderto answer this

ques-tion, we measured the synthesis of NA

protein

in infected cells. Viral proteinswere labeled with

[3Slcysteine

at 1, 3, 5,and 7 hpostinfection (Fig. SB).The celllysateswerethen analyzedon a10%polyacrylamideLaemmligel(7), and the

proteinswerequantitatedbyanAMBISradioanalytic imag-ingsystem(11). Synthesisofthe NAproteinof theNA/B-NS

transfectant viruswasmeasured and showntobetwotimes lower than that of wild-type virus at both 5 and 7 h postinfection. From this experiment,we canconclude that

3000

a -E C

a

CD

C M._

0 C)

4 Hours (p.i.)

8

VOL. 66,1992

on November 9, 2019 by guest

http://jvi.asm.org/

[image:5.612.69.557.72.511.2]

(6)

J. VIROL. B.

c., Z z m

en

;z

: z

0 HAO

HAl

fb40

WSNi'33 NA/B-NS

1 3 5 7 1 3 5 7 hours(p.i.)

3 P

:-;p..^* HA

PA N P .-NA

Actin

NSl

FIG. 5. Analysis of NA protein invirionsand in infected cells. (A) Western analysis of NA protein in virions. As described in Materials and Methods, a monoclonal antibody directed against carbohydratewasusedtoquantitatetheglycoprotein in the viruses. Theproteinsareindicatedby HAO(uncleaved HA),HAl (subunit 1

of HA), and NA at the right. (B) Viral proteins synthesized in infected cells. At different times postinfection (p.i.), the viral proteinswerelabeled with [35S]cysteinefor 30min and analyzedon

a10%acrylamide Laemmli gel. Thedifferent proteinsaremarkedat the right (see Fig. 2 legend). The position of the NA protein is

indicatedbyan arrow. TheamountofNAproteinwasdetermined

by counting the radioactivity in thegel in a radioanalytic AMBIS imagingsystem (11) with the NP and Ml proteinsascontrols.

the assembly of the NA protein into virions parallels its synthesis in infected cells.

DISCUSSION

Inapreviousreport, wedescribedatransfectantinfluenza

Avirus containing achimeric NAgenein which the

noncod-ing sequences are identical to those in the NS gene of

influenza B/Lee virus. This virus has many unique growth

characteristicsin tissue culture, and it is highly attenuatedin mice (14). Since such avirus could be aprototype for live influenzavirus vaccines, attemptsweremade tounderstand the molecular mechanism underlying these changed growth characteristics. Several lines of evidence obtained from theseexperimentssuggestthattheciselements derived from the influenza B/Lee virus gene are responsible for the

dramatic effects on transcription and replication of the

chimeric NAgeneof the NA/B-NS transfectantvirus. Itwas found that the NAgenehadasixfold-lower representationin

the purified viral preparation than did the remaining seven

RNAs. This strikingly lower representation ofone RNAis

compatible with the finding that the NA/B-NS transfectant virus has an approximately 5- to 10-fold-lower infectious particle-to-physical particle ratio than wild-type virus.

It is assumed that an infectious virus would require the

presence of a full complement of all eight influenza virus

RNA segments. Many of the NA/B-NS progeny virus,

however, lackanNAgene, sothatmore defective particles

areformed than is the casein awild-type virus infection. It

isnotclearwhetherthis5-to10-fold reduction in titer is only

areflection of the lower representation of the NA gene or

whether other factors also play a role. For example, some

viruses may contain defective interfering RNAs which would lower the infectivity titer of the preparation.

How-ever, in this context it should be noted that there is no

evidence for the presence ofsignificant quantities of defec-tive interfering RNAs in theNA/B-NSvirus.

The mRNA synthesis of the NA is also considerably

reduced in transfectant virus-infected cells. Surprisingly,

this does not lead to a commensurate reduction in protein synthesis. Both virus-infected cells and purified virus show

only a twofold-lower level of NA protein than wild-type virus-infected cells or purified wild-type virus itself. This higher-than-expected level of NA protein in thetransfectant virusmaybethe result ofagoodKozaksequencepresent in the chimeric NAgene. The chimeric NAgenehasanAin the -3positioninstead of the U found in thewild-typeNARNA (open triangle in Fig. 1). We believe that the twofold reduction in NAactivity doesnot significantlyinfluence the pathogenicity of thevirus, sincewe have constructed

trans-fectant viruses in which the expression of the NA gene is downregulated by a factor of 10 without affecting virus growth in tissue culture (3a). It should be noted that in the transfectant virus, expression of the othersevengenesis not

altered. Specifically, the NA/B-NS virus produces the same level ofHAproteinas found in wild-type virus, and compa-rable amounts of HAare packaged into the envelope ofthe virus. It appears that the attenuation characteristic is best

explained bythe lower synthesis ofNA-specific vRNA and

by the resultinglower representation ofparticles containing afull complement ofthe eight RNAsegments.

How influenza A virus packages its eight RNA genome segments remains an interesting question. In the past, two different mechanisms were proposed for the packaging of

influenza virus RNAs: one suggeststhat theeight RNAsare

selectively packaged, and the othersuggeststhatviral RNAs

arepackaged randomly (3, 8, 20). Evidence isnow accumu-lating to support the random-packaging mechanism. The

idea originated from the fact that influenza viruses have a

lowratio of infectious particlestophysical particles(3, 8). If

one assumes that an average of 11 RNAs are packagedper virion, the expected ratio is compatible with that found in

vivo (6). This modelwas also supported by the findingofa reassortant virus which contained two copies of the same segmentderived from twodifferentviruses(17), andfurther support for this theory came from a recent report which

described aninfluenza A virus which required nineRNAsin

order to be infectious (6). Our data in the present study

concerning the characterization of the NA/B-NS

transfec-tant virus also seem to favor the random-packaging

mecha-nism rather than the selective one. The lower level of

chimeric NA RNA found in virions is consistent with the

reduction of its synthesis in infected cells. In a selective

packaging model, it would be expected that approximately

equimolaramountsofchimeric NA RNA would bepackaged

into virus particles.

In summary, the present data suggest that influenza

vi-ruses for which the synthesis ofavRNA segment is down-regulated produce defective particles during replication.

Since the proteins of this virus areunalteredcompared with

those ofwild-type virus, attenuation must be the result of

inefficientcis-acting signals. We believethat thisprinciple of

attenuation canbe applied to otherviruses with segmented

genomes. For example, the introduction of modifications

into the noncodingsequences ofrotavirusgenes orofgenes

of other segmented double-stranded RNA viruses (16)

should also allow the pathogenicity of these viruses to be

altered. With the influenza virus system, we now plan to identify the nucleotide sequences in the NA/B-NS virus

which are responsible for the regulation of the synthesis of

vRNA and mRNA. This should help us to precisely

deter-A.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.63.294.74.238.2]

(7)

ATTENUATION OF CHIMERIC INFLUENZA A/B VIRUS 4685

mine the cis elements required for the replication of

influ-enzaviruses.

ACKNOWLEDGMENTS

This work wassupported byNational Institutes of Health grants AI-18998andAI-11823toP.P., a Max Kade Foundation fellowship toM.B.,and a NATO fellowship to A.G.-S.

REFERENCES

1. Brown, F. 1990.Introduction: modern approaches tovaccines. Semin. Virol. 1:1-2.

2. Centers for Disease Control. 1992.Outbreak of influenzaAina nursing home-NewYork,December1991-January 1992. Mor-bid. Mortal. Weekly Rep.41:129-131.

3. Compans, R. W., N. J. Dimmock, and H. Meier-Ewert. 1970. An electron-microscopic studyof theinfluenza virus-infectedcell, p.87-108. In R. D. BarryandB.W. J.Mahy(ed.), The biology oflarge RNA viruses.AcademicPress, NewYork.

3a.Doymaz,M., P. Palese et al.Unpublisheddata.

4. Enami, M., R. Fukuda, and A. Ishihama. 1985. Transcription and replication of eight RNA segments of influenza virus. Virology142:68-77.

5. Enami, M., W. Luytjes, M. Krystal, and P. Palese. 1990. Introduction ofsite-specific mutations into thegenomeof influ-enzavirus.Proc. Natl.Acad. Sci. USA 87:3802-3805. 6. Enami, M., G. Sharma, C. Benham, and P. Palese. 1991. An

influenza virus containing nine differentRNAsegments. Virol-ogy185:291-298.

7. Laemmli, U. K. 1970. Cleavageof structuralproteinsduringthe assembly of the head ofbacteriophage T4. Nature (London) 227:680-685.

8. Lamb, R., and P. W. Choppin. 1983. The gene structure and replicationofinfluenzavirus. Annu. Rev. Biochem. 52:467-506. 9. Luo, G., W. Luytjes, M. Enami, and P. Palese. 1991. The polyadenylation signal ofinfluenza virus RNA involves a stretch

of uridines followed by the RNA duplex of the panhandle structure.J. Virol.65:2861-2867.

10. Luo, G. X., and P. Palese. 1992. Genetic analysis ofinfluenza virus. Curr.Opinion Genet. Dev. 2:77-81.

10a.Luo, G. X., and P. Palese. Unpublished data.

11. Luo, G., and J. Taylor. 1990. Template switching by reverse transcriptaseduring DNA synthesis. J. Virol. 64:4321-4328. 12. Luytjes, W., M. Krystal, M. Enami, J. Parvin, and P. Palese.

1989.Amplification, expression and packagingof a foreign gene by influenza virus. Cell59:1107-1113.

13. Murphy, B. R., and R.G. Webster. 1990.Orthomyxoviruses,p. 1091-1152. In B. N.Fields andD. M.Knipe(ed.),Virology,2nd ed. Raven Press, New York.

14. Muster, T., E. K. Subbarao, M. Enami, B. P. Murphy, and P. Palese.1991. AninfluenzaAvirus containing influenzaBvirus 5' and 3' noncoding regions of the neuraminidase gene is attenuatedin mice. Proc.Natl.Acad. Sci. USA88:5177-5181. 15. Parvin,J. D., P. Palese, A. Honda, A.Ishihama,and M.Krystal. 1989. Promoteranalysisofinfluenzavirus RNApolymerase.J. Virol.63:5142-5152.

16. Roner, M. R., L. A. Sutphin, and W. K. Joklik. 1990. Reovirus RNAisinfectious. Virology 179:845-852.

17. Scholtissek, C., W. Rohde, E. Harms, R. Rott, M.Orlich, and C. B. Boschek. 1978. A possible partial heterozygote of an influenzaAvirus.Virology 89:506-516.

18. Seong, B. L., and G. G. Brownlee. 1992. A new method for reconstitutinginfluenza polymerase andRNA in vitro:astudy of the promoter elements for cRNA andvRNA synthesis in vitro and viral rescue in vivo. Virology186:247-260.

19. Shapiro, G.I.,T.Gurney, Jr., and R. M. Krug. 1987.Influenza virus gene expression: control mechanisms at early and late times of infection and nuclear-cytoplasmictransport of virus-specificRNAs. J. Virol.61:764-773.

20. Smith, G. L., and A.J. Hay. 1982.Replicationof the influenza virus genome.Virology 118:96-108.

VOL. 66,1992