Alphavirus nsP3 functions to form replication complexes transcribing negative-strand RNA.

(1)

0022-538X/94/$04.00+0

Alphavirus

nsP3 Functions

To Form

Replication Complexes

Transcribing Negative-Strand

RNA

YUN-FENWANG,t STANLEY G. SAWICKI, ANDDOROTHEA L. SAWICKI*

DepartmentofMicrobiology, MedicalCollege of Ohio, Toledo, Ohio43699

Received 8 March 1994/Accepted 30 June 1994

Thealphavirus mutant Sindbis virus HRts4, which has been assignedto the A complementation group,

possessed a selective defect in negative-strand synthesis that was similar although not identical to that observed for the Bcomplementationgroupmutanttsll(Y.-F. Wang,S. G.Sawicki,and D. L.Sawicki,J.Virol.

65:985-988, 1991).The causal mutationwasidentifiedasachangeofaC toaU residue at nucleotide 4903in thensP3openreadingframe thatpredictedachangeof Ala-268toVal.Thus,bothnsP3 and nsPlplayarole

selectivelyinthetranscriptionofnegative strandsearlyin infection. Theassignmentof the mutationcarried

byanAcomplementationgroupmutantof Sindbisvirus HR to nsP3wasunexpected,asmutationsin otherA

complementationgroupmutantsstudiedtodatemappedtonsP2.Another mutant withaconditionallylethal

mutation, ts7 ofthe Gcomplementation group, also possessed a causal mutation resultingfrom a single-residue change in nsP3. Negative-strand synthesis ceased more slowly after a shift to the nonpermissive temperatureints7- than in ts4-infectedcells,and ts7complemented tsll,butts4did not.However,the nsP3of bothts4 and ts7 allowed reactivation ofnegative-strand synthesis bystablereplication complexes containing

nsP4 from ts24.Therefore,mutations in nsP3 affectedonly earlyeventsinreplicationandprobably preventthe

formation and/or function of the initial replication complex that synthesizes its negative-strand template.

Becauseneither ts4norts7complemented10 Acomplementationgroupmutants, thegenesfor nsP2 andnsP3 functioninitiallyas asinglecistron. Weinterpretthesefindingsandpresentamodel tosuggestthat theinitial alphavirus replication complexisformed fromtightlyassociated nsP2 andnsP3, perhapsinthe form ofP23,

andproteolytically processedand trans-active nsP4 and nsPl.

Sindbisvirus,amember oftheAlphavirusgenusof thefamily Togaviridae, is an enveloped, positive-strand RNAvirus that replicates in bothvertebrate and invertebrate cells. The 49S genomeof the heat-resistant strain ofSindbis virus(SIN HR) is 11,703 nucleotides (nt) in length, exclusive of the cap and

poly(A)tract(48). After the virusenters thecytoplasmof the

infected cell and uncoats, thegenome is translated into two nonstructural protein (nsP) precursors,thepolyproteinsP123 and P1234, the latterby readthrough ofan opal termination codon at the end of the nsP3 reading frame. Readthrough occursatabouta20%efficiency, resultinginthe underproduc-tion of nsP4polypeptides comparedwithnsPl, nsP2,and nsP3

(27). The P123 and P1234 polyproteins are proteolytically cleavedbyapapain-likeprotease activityinnsP2toyield the

maturensPl,nsP2, nsP3, and nsP4 orthe fusion proteinP34, which are numbered according to theirgene order (7, 8, 18, 27). The nsP4 sequence contains the conserved GDD motif and functions as the viral polymerase (2, 12, 24, 25). The pattern ofprocessing of the polyproteins changes at different

times ininfection, giving risetodifferent intermediate

polypro-teins(7, 17, 44). Once formed, the initialreplication complexes

synthesize genome-length negative strands that become

tem-plates for synthesis of the genome RNA and of a 26S sub-genomic mRNA that is translated into the viral structural

proteins.

During the alphavirusreplication cycle, negative-strand

syn-thesis occurs only at early times and requires simultaneous viralprotein synthesis, whereaspositive-strand synthesis, once

*Corresponding author. Phone: (419) 381-4337or(419) 381-3921.

Fax: (419)381-3002.

tPresentaddress: Laboratory ofVirologyand Parasitology, Lind-sleyF. Kimball Research Institute of The New YorkBlood Center, NewYork, NY 10021.

started, is stable and continues even if protein synthesis is terminated (36, 40). Thus, therearetwoformsoractivities of the alphavirus RNA-dependent RNA polymerase. We (36)

hadspeculatedthatnegative-strand synthesiswasregulated by amechanismutilizingoneor moreshort-lived viralnsPs,which were modified by proteolytic cleavage, or some other

post-translational activity. For instance,uncleaved orincompletely cleaved precursors of the nsPs would function in alphavirus negative-strandRNAsynthesisbutaftercleavage would form

the stable positive-strand polymerase. Experimental evidence

supporting this notion has beenobtainedrecently (25, 45).

We had studied the requirements for negative-strand syn-thesis with RNA-negative mutantsofSIN HR that had been screened(40)fortherapidandselective loss ofnegative-strand synthesis after ashift to40°C. Mutants ts4, ts6, and tsll had

been identified(40) asdefectiveincontinuing negative-strand synthesis under such conditions, although only ts4 and ts1l were defective specifically in negative-strand synthesis (37).

The temperature-sensitive (ts) defect in tsl1 was mapped to nsPl (16, 49, 50), a protein that also possesses methyltrans-ferase and guanylyltransferase activities (28). ts4, which had

beenassignedtosubgroupIIof the Acomplementationgroup (37), was of interest because, in addition to its being an RNA-negative mutant, it expressed a temperature-indepen-dentreduction inoverallRNAsynthesis and in thesynthesis of

subgenomic 26SmRNArelativeto49SgenomeRNA(20, 37). Although the mutations of the subgroupII A complementation

groupmutantshavenotbeen mapped, subgroupImutantsthat aredefective in polyprotein cleavage and 26SRNAsynthesis

carrymutations in nsP2 (16, 39). Therefore, weundertook a

detailedcharacterization of ts4.

(Someof theresultswerepresentedatthe 1992meeting of

the American SocietyforVirology.)

6466

on November 9, 2019 by guest

http://jvi.asm.org/

(2)

MATERMILS AND METHODS

Viruses and cells. The mutants ts4, ts7, and tsll of the heat-resistant strain ofSindbis virus (SIN HR) were isolated

originally by Burge and Pfefferkorn (3, 4). A sample of ts4,

obtained fromEllen Strauss,wasplaque purified, and a stock

was prepared following growth at 30°C in chicken embryo

fibroblast (CEF)cells at amultiplicity of infection (MOI) of

0.1. Arevertantof ts4 (ts4R) wasisolatedfromaplaque that formedat40°Candwasplaque purifiedonce at30°C and once

at40°Cbeforegrowth ofastockat40°C. SIN ts7 is a complex

mutant with two conditionally lethal mutations that affected

plaqueformation: achangeofGtoAat nt3243, predicting a

changeofAsp-522toAsninnsP2,andachangeof Uto Cat

nt 5035, predictinga change ofPhe-312 to Ser in nsP3 (16);

onlythe Phe-312toSerchangein nsP3 ledto anRNA-negative

phenotype. TotollOl:ts7B5 is a hybrid virus containing only

thensP3 mutation(16).Itwasproduced fromconstructsof the

infectious SINcDNAclone(33)inwhich thets7regionfrom nt

4845 to nt 5262 replaced the corresponding parental

pTotollO1 sequence.

Synthesisof cDNAandconstruction ofrecombinant viruses.

The infectious SIN cDNA clonepTotollO1 wasused to create three deletion mutants of the parental sequence to enable

rapidselection ofrecombinants. TEB14 is the deletion vector

ofpTotollO1 lackingtheregionfromnt 1406to nt2288after

digestionwith Eco47III andBglII;TW20 lacks the region from

nt 2713 to nt 4280 after digestion with ClaI and AvrII; and TSH6 lacks theregion fromnt5114to nt 6917after digestion

with SfiI andHpaI (see Fig. 2). Unless otherwise indicated,

restriction and other enzymes were obtained from New

En-gland Biolabs (Beverly, Mass.). To generate the deletion

vectors, pTotollO1 cDNAwas digested with pairs of

restric-tion endonucleases, and thefragmentswereblunt ended with

the Klenow fragment of Escherichia coli DNA polymerase

(U.S.

Biochemicals, Cleveland, Ohio) at 37°C for 15 min or

withbacteriophage T4 DNApolymerase (U.S. Biochemicals)

at12°C for15 min inbuffercontaining80,uM deoxynucleoside

triphosphates, purified by electrophoresis on

low-melting-pointagarosegels

(FMC

BioProducts, Rockland, Maine),and

self-ligated

with T4 DNA

ligase

(Promega,

Madison, Wis.) in the presenceof 50 ,uMATP-20 mM dithiothreitolat 16°C for

4hbefore transformation into MC1061cells.

To construct recombinant viruses containing the various partsofthe nsP genes from SINts4,

purified

ts4 virion RNA

was reverse transcribed withMoloney murine leukemia virus RNase H-reverse

transcriptase

(GIBCO

BRL, Grand

Island,

N.Y.)

inareaction buffer

containing

50mMTris-HCl

(pH 8.3),

75 mM KCl, 3 mM MgCl2, and 20 mM dithiothreitol by incubationat42°Cfor1h. Toobtainafirst-strandcDNA copy of thensPl-nsP2

region, oligonucleotide

RH15-2

(nt

2811to nt

2825;thegiftof J.

Strauss)

wasusedasthe first-strandprimer;

to obtain a first-strand nsP2-nsP3 cDNA,

oligonucleotide

YW15.6

(nt

5348 to nt

5366)

was used as the first-strand

primer. One-twentieth of the first-strand cDNA was then

amplified by

a PCR method modified for

amplification

of cDNA

products larger

than 1kb

(14, 51).

Asshown in

Fig. 2,

YH47, a positive-sense oligonucleotide

containing

the SstI

restrictionsite,theSP6promoter sequence,and the5' 15ntof theSindbis genomewasusedasthesecond-strand

primer

with

oligonucleotide

RH15-2 to

amplify

the nsP1-nsP2 sequence, and

oligonucleotide

ID2

(nt

1366to nt

1380, positive

sense)

was used as the second-strand

primer

with

oligonucleotide

YW15-6 to

amplify

the nsP2-nsP3 sequence. The reactions

were carried out with

AmpliTaq

DNA

polymerase

(Perkin

ElmerCetus,Norwalk,

Conn.)

in10

IJl

ofasolution

containing

lx PCR buffer (50 mM KCl, 10 mM Tris [pH 8.3], 2.5 mM

MgCl2, 0.1% gelatin)and the four deoxyribonucleotides, each

at aconcentrationof 0.2 mM. The sample was heated for 1 min

at95°C before beginning the amplification cycle (1 min at 95°C for denaturation, 1 min at 43°C for annealing, and 5 min at 72°C for polymerization) for a total of 20 cycles followed by a final extension cycle of 30 min at 72°C. The products were electrophoresed through 0.8% low-melting-point agarose gels

andpurifiedwitheitherphenol-chloroform or the GeneClean

procedure (GeneClean II; Bio 101, Inc., La Jolla, Calif.). Following digestion withSstI(GIBCO BRL) and ClaI or with

BglIIandSpeI and repurification of the fragments, the

nsPl-nsP2and nsP2-nsP3 ts4 cDNAs were ligated into the TEB14 or TW20 deletion vector, respectively, in place of the parental SIN sequence to create the hybrid clone pTotollOl:ts4SC, which contained nt 1 to nt 2713 of ts4, encoding nsPl and the N-terminal part of nsP2 coding region, and pTotollOl:ts4BS, which contained nt 2288 to nt 5262 of ts4, encoding the C-terminal half of nsP2 and the N-terminal half ofnsP3. For

subcloningofpTotollOl:ts4BS, the cDNA was digested with

restriction enzymes that divided the sequence into two parts before ligationinto the deletionvectorTEB14 orTSH6 (see

Fig.2). Thus,the clonepTotollOl:ts4BA contained the region

from the BglII to AvrII site (nt 2288 to nt 4280), and

pTotollOl:ts4AScontained the region from theAvrII toSpeI

site (nt4280to nt5262).

The doublemutantTotollOl:ts4:24R1 was created by

com-bining inthesame infectious pTotollOl cDNA clone the ts4

nsP3 mutation contained in the SstI (nt -90)-SpeI (nt 5262)

fragment of pTotollOl:ts4AS and the ts24Rl nsP4 mutation

that reactivated negative-strand synthesis at 40°C and was contained in the SpeI-SstI fragment from pTotollOl:24R1

(35).This strategywaspossiblebecause the nsPl-nsP3 (ts4)or

nsPl-nsP4 (ts24Rl) coding sequences of each genome had

been assayed for phenotype, and regions other than those

containing each of the two mutations gavewild-type

pheno-types (35; this study). If there were any other undetected sequence differences between the mutant and SIN HR or

TotollOl, such changes did not lead to ts virus production.

Likewise, the double mutant TotollOl:ts4:11A1 contained

boththe ts4 nsP3 mutation and the tsll nsPl mutation andwas

createdby ligatingtheClaI(nt 2713)-AatII (nt 7999) fragment

of pTotollOl:ts4AS in place of the parental sequence in

pTotollOl:tsllAl,which contains thets1lnsPl mutationat nt

1101

(16, 50).

The presence of both mutations in eachdouble

mutantcDNAwasconfirmedby sequencingthe cDNAclones and,for TotollOl:ts4:24R1 virus, by expressionofnew

condi-tionallylethalphenotypes comparedwith the TotollOl:24R1

virus.

Sequencingandanalysis ofpredicted proteinchanges. The

hybridcDNAwassequenced bythedideoxychain termination

method (34)withSequenase2.0

(U.S.

Biochemicals).

The ts4 and ts4R virion RNAwas

sequenced by

the

dideoxy

method with avian myeloblastosis virus reverse

transcriptase

(Life

Sciences,St.

Petersburg,

Fla.).

The

parental

and the

predicted

mutantnsP3 sequenceswere

analyzed

for

secondary

structural features (5, 10,

11),

including

folding,

antigenicity,

and

hy-drophobicity, with LaserGene sequence

analysis

software

(DNASTAR, Inc., Madison,

Wis.).

Invitrotranscription andtransfectionofCEFmonolayers.

The plasmid cDNAs were linearized with XhoI restriction

endonuclease before incubation withSP6RNA

polymerase

for the

preparation

of

transcripts

that were used

immediately

thereafter for transfection of CEF

by

the DEAE-dextran method

(33).

Infected cellswere maintainedat

30°C

in Dul-becco'smodified

Eagle's

minimal essential medium

(DMEM)

on November 9, 2019 by guest

http://jvi.asm.org/

(3)

containing 5% fetal bovine serum until the development of

cytopathic effect at about 36 to 48 h postinfection

(p.i.),

at

which time thevirus-containing mediumwasharvested. Infection and viral RNA labeling. CEF monolayers in 35-mmor60-mmpetridisheswereinfected with virus(MOIof

100)at30°C.

[3H]uridine

(50 ,uCi/mlfor overall RNAsynthesis

or200,uCi/mlfornegative-strandRNAsynthesis; ICN, Irvine,

Calif.) incorporationinto viral RNAwasdeterminedby pulse

labeling in the presence of 20 ,ug ofdactinomycin per ml at

differenttimes after infection and countingthe acid-insoluble

radioactivityinafraction of the cellextracts.Where indicated

inthe text, infected cellswereshiftedto40°C duringtheearly

phaseofinfection,at atime when about10%ofthe maximum

rate of RNA synthesis had been reached at 30°C, or were

incubatedcontinuously at30or40°C.Cultures maintained at 30°C were labeled for 1-h periods beginning 1 h before

duplicate cultures were shifted to 40°C. Cultures shifted to

40°Cwere pulse labeled for 30-min

periods beginning

at the

time of the shift.

Determination ofnegative-strandRNAsynthesis. Negative-strand RNA synthesiswas determined by quantifying labeled

negative-strand RNAin purified replicativeform

(RF)

RNA,

obtained by RNase A treatment of the

[3H]uridine-labeled

viral replicative intermediates (RIs). The RF RNAwas heat denatured and hybridized in the presence of an excess of

unlabeled, purified 49S virion positive-strand RNA

(41).

A

value of50%indicates that50%of thetotal labeledRNAwas

negative-strand RNA and 50% was positive-strand RNA.

Because the RF is derived from RIs that are also active in

positive strand synthesis, a value of 50% labeled

negative-strandRNAalso indicates that100%of the total viralnegative strands in these RIswere

newly synthesized during

the

pulse

period.

Complementation. Complementation tests were performed

asdescribedbyStraussetal.

(46)

andmodifiedbySawicki and Sawicki

(37). Briefly,

infected

monolayers

werewashed with 40°Cmedium once at1 h

p.i.

and

again

forthree timesat4h p.i. to remove unadsorbed virus. Virus released at 40°C was

harvested at 7 h p.i. and quantified by plaque assay at 30°C.

Thecomplementationindexwasthe ratio of theyieldof virus

frommixedlyinfected CEFcells, usingeach virusat anMOI of

20,at40°Cdividedbythesumof theyieldsfrom cells infected at anMOI of 20byeach mutantalone.

RESULTS

SINts4hadaselectivetsdefectinnegative-strand synthesis.

Normally, alphavirus negative-strand RNA synthesis occurs

withinthe first 5to6 h p.i. at30°C and ceasesthereafter. We

confirmed that in ts4-infected cells, negative-strand synthesis

ceasedrapidly and within 30 min aftera shiftto 40°Cat any

time during the normal period of negative-strand synthesis.

Thisobservationis consistent withts4'spossessing a ts defect

thatselectivelyaffects theproductionofnegative strands(Fig.

1A). This was in contrast to positive-strand synthesis, which continued at 40°C but, as observed for all RNA-negative mutantsofSindbisvirus,neverincreasedtothe maximal rate observedat30°C (Fig. 1B). Revertants of ts4 (ts4R) occurred

at afrequencyof1 x 10-5to 6 x

10-',

which was consistent

with reversion ofa single point mutation, and were found to

lackthetsdefect in negative-strandsynthesis(data not shown).

ts4mutanthad asinglemutation in the nsP3coding region.

Although the reversion frequency of ts4 predicted a single

point

mutation,wetookinto consideration the possibility that

ts4 possessed mutations in both nsPl and nsP2 because the

complementation

patternof ts4was complex (16, 37).

More-A

40 35

* 30

1 25

! 20

, 15 10

2

B

1104

E

06103

z

1.5 3.0 3.5 4.0 4.5 5.0 5.5

shitt to40C

1 23 4 5 67 8 9

Hoursp.i.

FIG. 1. Kinetics ofts4 RNAsynthesis. (A) Negative-strand synthe-sis. SINts4-infected CEFmonolayersweremaintainedat30°C(0)or shiftedto40°Cat3hp.i.(@)and incubated inmedium thatcontained

200 ,uCiof[3H]uridine(3H-Ur)and 20pugofdactinomycinperml at therespectivetemperaturesfor30-minperiodsat40°Corfor 60-min periods at 300C. Cultures were harvested at the end of the pulse period, andtheircontentof labelednegative-strandRNAwas

quan-titated as described before (40). (B) Positive-strand synthesis. SIN

ts4-infectedcellsweremaintained at30°C(0) orshiftedfrom30 to

40°C at3 or4h p.i. (0),pulse labeledwith [3H]uridinefor30-min periodsat400Corfor 60-minperiodsat30°C,harvestedatthe endof

the pulse period, and analyzed for total acid-insoluble radioactivity, greaterthan95% of whichwasin49S and 26Spositive-strandRNAs.

over,the assignment of ts4 tothe Acomplementation group

didnotruleoutthepossibilityof itshavingamutationinnsP3,

becausecomplementationbetweeneither oftwonsP2mutants

belonging to the A complementation group and the nsP3 mutantts7 of the G complementation group, while expected from earlierresults, was notobserved (16) and becausetsl8, assigned originally to the Gcomplementation group, turned

out to be an nsP2 mutant (16). Therefore, we initially

con-structedtworecombinant infectious cDNA clonesto screenfor mutations innsPl,nsP2, and the conserved regionof nsP3 of ts4. Onecontained all of thecodingsequencesofnsPl andan

N-terminal portion of nsP2 from ts4 (pTotollOl:ts4SC; the SstI-ClaIfragment);the other contained theremaining portion of thecodingsequencesofnsP2, only slightly overlappingthe

pTotollOl:ts4SCsequences,andthecodingsequencesfor the

conserved region of nsP3 (pTotollOl:ts4BS; the BglII-SpeI

fragment)inplace of the corresponding regions of the parental

cDNAclonepTotol101(Fig. 2).Use ofthe uniqueSpeI siteat nt 5262 facilitated cloning and allowed assessment of the conserved N-terminalportionof the nsP3 sequence(nt4101 to

5747) (48). Because the C-terminal portion of nsP3 can be

deleted withoutsignificantly affectingtranscription(23),it was

unlikely that the conditionally lethal mutation of ts4 would

reside in this part of nsP3. Following transcription and trans-fection of the recombinant RNA intoCEFcells, the resulting

hybrid viruses were analyzed for ts growth, for reversion

frequency, and for their ability to continue negative-strand

synthesis afterashift to40°C early in infection.

As shown in Table 1, only TotollOl:ts4BS, which is the

hybrid virus that contained the BglII-SpeI region, was ts for

growth, for virus production, and for negative-strand synthesis

(i.e., it ceased negative-strand synthesis within 30 min after a

shift to 400C early in infection). The BglII-SpeI region was subcloned to identify the region containing the mutation. TotollOl:ts4BA (BglII-AvrII)had a wild-type phenotype, but

TotollOl:ts4AS (AvrII-SpeI)was ts for virus production and

fornegative-strandsynthesis(Table 1). Thus, the ts4 mutation

was locatedbetweennt 4280 andnt 5262 and was within the

openreadingframe for the conservedportion of nsP3.

Two changes from the parental TotollO1 sequence were

on November 9, 2019 by guest

http://jvi.asm.org/

(4)

-RH15-2 (2811)

.YW15.6(5348)

YH47(SstI-Sp6.lS)

+IDI(1380)

5I

I nsP3 I

nsP4

Eco47m BglII 6aI

1406 2288 2713

AvrII SflIEJpeI 4280 5114 5262

Hpal 6917

11

3'

AatII 7999

] TotolIOl:ts4BS

Totoll0l:ts4SC

Totol IOl:ts4BA

]

Totol101:ts4AS

FIG. 2. Strategy for mapping the responsible mutation ints4.Infectious cDNA clones that containedpartsof thets4 nsPcodingregionsinplace

ofthecorresponding regions of parental pTotollO1 weremadeasdescribed inMaterials and Methods. Unique restriction endonucleases andtheir

cleavagesitesaremarked under the schematic of the viralnsPlthroughnsP4coding regions; primers and deletionvectorsareindicatedabovethe

schematic. The hatched regionsrepresentsequencesfromts4;theopenregionsarethose from pTotollOl.

found when the region from nt 4280 to nt 5262 of TotollO1: ts4ASwassequenced (Fig. 3). One was a change of a C to a U residue at nt4903 that would change the Ala-268 in nsP3 to a Val; the secondwasachange ofaU to aCresidue at nt 5060 that would not alter the coding of the Val-320 residue and would be silent (Fig. 3). The same two base changes were presentin theoriginal ts4 genomic RNA purified from virions

(datanotshown).The ts4R genomeretained the silent change

at nt5060 buthad theparental C residue atnt 4903 andwas

thusatrue revertant.Asshown inFig.4, thepredictedchange of Ala-268to Val in ts4wasinaregionof nsP3 that is highly

TABLE 1. Analysisof ts4 recombinant viruses

Fragment Efficiency ofplating

Negative-Virus replaced Growtha (titerratio, strand

(nt) 40°C/30°C) synthesisb

Totoll0l wt wt

ts4 ts 1 x10-5_6x10-5 ts

TotollOl:ts4SC 1-2713 wt 1.9x 10-1 wt

TotollOl:ts4BS 2288-5262 ts 2.0x10-5 ts

TotollOl:ts4BA 2288-4280 wt 1.2x10-1 wt

TotollOl:ts4AS 4280-5262 ts 2.3x10-5 ts

awt,wildtype; ts,recombinantsshowingtsgrowthorarapid,

temperature-dependent cessation ofnegative-strand synthesis that mimicked that ofthe mutantts4.

bThetemperaturesensitivity ofnegativestrandsynthesiswasdeterminedby shiftinginfected culturesto40°Cat3 hp.i.orwhen therateofpositive-strand synthesiswas 10% ofthe maximal rate and labelingwith [3H]uridine (200

±Ci/ml)for30-minperiodsbeforeharvesting.Thecontentoflabeled

negative-strand RNA in theisolatedRFswasdeterminedasdescribedinMaterials and

Methods.

conservedamongalphaviruses (12, 21,43,47). TheotherSIN HRts RNA-negativensP3mutantis ts7of the G

complemen-tation group,whoseconditionallylethalmutation isachange ofPhe-312toSer(16).The altered ts7residueislocatedwithin the same highly conserved N-terminal half of nsP3 but 44

conseved domain

S.

4280

SIN HR

Totol101:t4AS

ts4R

TotollOl:ts7BS

--I-. 5747

Sp6

5262 268 312 320

Ala Ph. Val

INs

4903 5035 S560

C U U

268 320

Val Val

4903 s506

U C

268 320

Vl oval

4903 5060

C C

312

9r

5035 C

FIG. 3. Relativemappositionsofthepredictedamino acidchanges innsP3 ofToto:ts4AS, ts4R,andToto:ts7B5 comparedwiththose in SIN HR and TotollOl. The data for Toto:ts7B5 were taken from Hahnetal.(16).

.

nsPl

SstI

nsPZ

m I 0.01--a I I I 0 W.Ps 0

u a . 0 s

I

.1-I

i---I

0

I

r

on November 9, 2019 by guest

http://jvi.asm.org/

(5)

260 270 280 290 300

SIN HR Z0, rrATarLCM sAMTFrssvnnHRLNr vaKEVTlVC;S;STFwLaranalNVKVQ; NrnrrvrAHT SIN Ockelbo 257 ---______________________________Y_---SIN t84 257 ---_---

a

---SIN ts7 257

---E

---SFV 256 --R-V----R----A--IA----HQ--SNV----F----YHVDG----K-E--L--D-TV-SV-SP RRV 256 ----V----R----A---A---K--T-AII----F----YR-EG----K-DR-LI-DQTV-SL-SP

ONN 256 ----V----R---A---N-HTTSII----F----Y--EG----K-S-AL--DHNV-SR-SP

HID 254 --C-V----R---AAQ--QF---F----Y--PG--R-A-SA-M---HDV--L-SP VBE 261 --S---IH---Q--KASRPEQI---F----YR-TG---I--SQPI--S-KV--YIHP

FIG. 4. Comparative aminoacidsequencesofalphaviralnsP3proteinsin theregionof the ts4and ts7 substitutions. Thechangesin SINts4and SINts7areoutlined.Sequencedata: SIN HR from Straussetal.(48);Ockelbofrom Shirako et al.(43);Semliki Forest virus(SFV)fromKeranen andKafiriainen (20);Ross River(RRV), O'Nyong-nyong(ONN),andMiddleburg(MID)virusesfrom Strauss et al.(47);and Venezuelanequine encephalitis (VEE) virus, Kinneyetal. (21).

codons downstream of the altered residue in ts4(Fig. 4). We

compared the phenotypes of the two nsP3 mutantsto

deter-minewhether the two residueswere part of the same func-tional domain.

Comparativeanalysisof SIN ts4 and ts7. In additionto two

mutations in the nsP3 coding region, only one ofwhich was conditionally lethal, theoriginalts7mutant has a third muta-tionat nt3247 in the nsP2coding regionthat conferstsplaque

formation butanRNA+/-phenotype (16).Therefore,weused

the hybridvirus TotollOl:ts7B5(16),which containedonlythe

conditionally lethal U-to-C change at nt 5035 in the nsP3 coding region (Fig. 3). We confirmed that our stock of TotollOl:ts7B5wastsfor overall RNAsynthesisand measured

itsefficiencyofplating (thetiterat40°Cdividedbythe titerat 30°C) as 7 x 10-5. This was necessary to ensure that the

phenotype that we observed was not due to leakiness, as reported (16) for the ts7B3 recombinant that contained both nsP3 base changesand had areversion frequency of 10-2 to 10-3.When TotollOl:ts7B5-infected cellswereshiftedto40°C earlyin infection, positive-strand synthesiscontinued but did

notincreasesignificantlyabove therate atthe time of the shift

(datanotshown). Thus,TotollOl:ts7B5replication complexes

formedat30°Cwerestable and functionedat40°C. However,

such cultures prematurely shut off negative-strand synthesis (Fig. 5). The cessation of negative-strand synthesis in Toto 1101:ts7B5 occurred sooner than the normal cessation ob-served at the end ofthe early phase of infection in parental

SIN HR-orts4R-infectedcells,neither of whichwasdefective in negative-strand synthesis, but it was not as rapid as that observed for tsll and ts4, two mutants that are selectively defective in negative-strand synthesis. Themore rapid cessa-tion ofnegative-strand synthesis after the shifttothe nonper-missivetemperature bytsll and ts4 than by ts7B5was repro-ducible and suggested the possibility that negative-strand synthesis was immediately inhibited in tsll- and ts4-infected cells.

Mutations of nsP3 afect the initial but not the stable

replication complex. To identify the point in the pathway of

negative-strand RNAsynthesisatwhichts4 nsP3wasdefective,

we monitored events required for formation of the initial

replication complex active in negative-strand synthesis sepa-ratelyfrom thosenecessaryfor the transcription of negative-strand RNA once the stable replication complex had been formed. Thiswaspossible becauseamutant nsP4 encoded by thegenomesof SIN ts24 and itsrevertants24R1 and 24R2 led tothereactivationat40°C of negative-strandsynthesis by viral

replication complexes that had formedearlier andwereactive

in positive-strand RNA synthesis at the time of the shift to

40°C (35, 42). We constructed double recombinant viral

ge-nomesthatexpressed both the ts4 nsP3 mutation and thets24

nsP4mutation, i.e.,TotollOl:ts4:24R1, and in thesame

exper-imentscompared this virus with a double recombinant virus expressingthe tsll nsPl mutation and the ts24 nsP4 mutation

(Toto1O1:tsll:24R1). Cells infected with either a double

mutant, an individual mutant, or parental TotollO1 were shifted to 40°Cat 6 h p.i. and incubated for afurther2 h at 40°C in the presence of cycloheximide to prevent further

proteinsynthesis and of[3H]uridinetolabelnewly synthesized

RNA. The viral RIs, which contain allof the viral

negative-strandRNAs,wereconvertedtoRFsbyRNasetreatment, and theircontent ofnewly synthesized negative-strand RNAwas measured. Cells infected with 24R1 virus reactivated

negative-strand synthesis at 40°C and had 5 to 13 times more labeled

negative strands than cells infected with only the ts4 ortsll mutantor withTotollO1 (Table 2). Although the individual

mutantsexpressing mutantnsPl ornsP3 proteinceased

neg-ative-strandsynthesis similarly earlyininfection,theirabilityto

reactivate negative-strand synthesis late in infection differed

dramatically.The nsP3 doublemutantTotollOl:ts4:24R1fully

reactivated negative-strand RNAsynthesis similarlyto 24R1,

whereas the nsP1 double mutantToto:tsll:24R1 wasblocked in reactivation (Table 2).

We alsoanalyzedtheindependentlyisolated double

recom-a

50

.4

co 0

C C

;~20

z

LL 10

10

n 0 0.5 1 1.5

Hours after shift to 400C

FIG. 5. Kinetics of cessationof negative-strand synthesis. Cultures wereinfectedat anMOI of 100at30°C withnsPl mutanttsll (0), nsP3 mutantts4(A), TotollOl:ts7B5 (-),orrevertantts4R(0) and maintainedat30°C until 3.5 h p.i., when theywereshiftedto40°C. The continuedsynthesis of viral negative strandsat40°Cwasmonitored by determining the amountofradiolabeled RNA in purified replicative

structures(RFs)in30-min pulse periods begunatthe time of the shift

to40°CasdescribedinMaterials andMethods. Theamountofnewly synthesized negative-strand RNA isexpressedas apercentageof the total labeledRF RNA innegative strands.

310 320

q5=0

#V-.

on November 9, 2019 by guest

http://jvi.asm.org/

(6)

TABLE 2. Comparative analysisof theabilityof nsP3, nsPl, and nsP2mutants toeffectreactivation ofnegative-strand synthesis

Exptno. Virus Mutant nsP Negative-strandRNAc

1 24R1 nsP4 2,010

Totol101 439

ts4 nsP3 395

Toto:ts4:24Rl nsP3+nsP4 1,810

tsll nsPl 155

Toto:tsll:24R1 nsPl+nsP4 501

2 24R1 nsP4 2,590

Totol101 428

ts7 nsP2+nsP3 1,663

Toto:ts7B5-3 nsP3 462

Toto:ts7B5-7 nsP3 551

Toto:ts7B5:24Rl nsP3+nsP4 1,395 Toto:ts7B5:24R1 nsP3+nsP4 1,543

aVirus-infectedcellswereshiftedto40°Cat6 hp.i. and labeledat40°C with [3H]uridinefrom 6to8 hp.i.inthepresenceofcycloheximide.Viruses fromtwo

independent plaques of Toto:ts7B5 and from twoseparate cDNA clones of

Toto:ts7B5:24R1 were tested. The viral RFs were obtained as described in

Materials and Methods. The valuesaretheaveragesofthe RF RNAcontentof

labelednegative strands fromtwotofourduplicate cultures for each virus.

binantviruses containing the ts7 nsP3 mutation and the ts24 nsP4mutation, i.e.,TotollOl:ts7B5:24R1.Thisviruswasfound to be similar to TotollOl:ts4:24R1 and reactivated negative-strand synthesis (Table 2). Thus, nsPl functioned in the transcriptionofnegativestrandsbypreformedorstable repli-cationcomplexes aswellasby newlycreatedorinitial replica-tioncomplexes. Furthermore, thedefect innsPlpossessed by

ts1lwasexpressed bymatureand stablensP1 proteins synthe-sized hoursearlier andactive inpositive-strand synthesisatthe time of theshift.On the otherhand,thetwomutations innsP3

which were found in ts4 and ts7 prevented negative-strand synthesis earlybut didnot affect thetranscription ofnegative strands by preformed complexes late in infection. Because thesetwo mutations in nsP3 affectedonly early functions,we interpret these results to mean that they affected only the initial replication complex during or immediately after its

formation and did notaffect theabilityof thestablereplication complextorecognizethepromoter fornegative-strand synthe-sis orthe actualpolymerizationof the negativestrands

them-selves.

Although the altered nsP3 of ts7 in Totoll0l:ts7B5 lost

negative-strand synthetic activity at40°C (Fig. 5), theoriginal ts7reactivatednegative-strand synthesisat40°C byitself and in

the absence of the 24R1 mutation(Table 2). Apparentlythis is

due to the conditionally lethal mutation in nsP2 of ts7 that predictsachangeofAsp-522toAsn in theCregionofnsP2,a region already implicatedinsubgenomicmRNAsynthesisand polyprotein cleavage (16). Thisbringsto three the number of

separatetsalterations in the CregionofnsP2(Ala-517to Thr [tsl7], Asp-522 to Asn [ts7],andAsn-700 to Lys [ts133] [39]) that lead to reactivation of negative-strand synthesis at 40°C,

two of which (tsl7 and ts133) also confer ts 26S mRNA

synthesis. The alteration ofAsp-522 to Asn led to a level of negative-strand synthesis similar to that observed for each of the othertwonsP2 mutants and half the level observed for the 24R1 nsP4 mutation (39). We interpret these findings to

suggest that the nsP2 protein functions in some unknown

manner, either directlyorpossibly indirectly, by affecting the

activityof other essentialfactors,to influencepromoter recog-nition for transcription of negative-strand versus positive-strand RNAby stablereplication complexes.

TABLE 3. Complementation analysis Virus" Mutant Complementation

index'

lrusprotein

tsl8

ts4 ts7 ts7B5 ts6

tsll(B) nsPl 5.20 1.60 3.00 5.2 76

tsl5 (Al) 0.61 0.03

tsl7(AI) nsP2 0.04 0.60 0.22 1.9 300

tsl8 (G) nsP2 0.24 0.05

ts2l (AI) nsP2 0.12 0.37

ts24(Al) nsP2 0.63 0.23 ts133 (Al) nsP2 0.63 0.04

tsl4(AII) 0.71 0.06

tsl6 (All) 1.08 0.38

tsl9(AII) 0.07 0.19

ts138 (AII) 0.07 0.18

ts4(All) nsP3 0.80 0.30 19.8

ts7(G) nsP3 0.12

ts6 (F) nsP4 2.40 19.80 24.8

aThecomplementation groupto which each mutant was originally assigned (4,

46)is shown inparentheses.The two subgroups into which the A complemen-tation group mutants wereplaced subsequentlyare also indicated (37).

b Complementationvalues arethe yieldof themixedinfections divided by the sumoftheyieldsof theinfectionswith each mutant alone. Values of less than 2 are taken to indicate that complementation did not occur. Analysis was

performedasdescribedby Sawickiand Sawicki (37).

Wealsodetermined whether both the tsl 1 nsP1 and ts4 nsP3 mutationswerecompatible in the same genome and whether their effectswereadditive. The presence of both mutations in the recombinantclones was verified by sequencing the double recombinant TotollOl:ts4:tsllAl cDNA (data not shown). Transcription of the TotollOl:ts4:tsllAl cDNA and transfec-tion of the resulting RNA gave virus that had an efficiency of plating of less than 10-8, consistent with the presence of two mutations. TheTotollOl:ts4:tsllA1 virus continued positive-strandsynthesisafterashift of theinfected cells to 40°C early ininfection but abruptly ceased negative-strand synthesis (data

notshown). Thus,both mutations weretolerated in the same

genome, and replication complexes containing both mutant

proteins did not show detectably greater defectiveness than

observed for replication complexes containing each mutant

protein separately.

Complementationanalysis of nsP3 andnsP2 mutants.

Be-cause intracistronic complementation was reported to occur

with SIN mutants (16),we tested theability of thetwo nsP3

mutantsts4 andts7, which had been assignedtothe A and G

complementation groups (4, 46), respectively, to rescue each

other and to complement mutations mapped to other nsPs.

Complementation analyses(Table 3) showed that ts4

comple-mented the nsP4mutantts6butnot mutantsencoding

defec-tivensPl, nsP2,ornsP3 proteins. Thesefindingswere

consis-tentwith andextendedour

findings (37).

Because ts7 contained a second ts mutation in nsP2

(16)

and showed a minor defect in processing of P123 and P1234 polyproteins uponashiftto40°C,leadingto someaccumulation of P23

(17),

the nsP3 recombinant Toto:ts7B5 and ts7 were tested. Toto

1101:ts7B5 did not complement ts4 but did

complement

the

nsP4 mutant ts6 and also the nsPl mutant tsll,

although

weakly. However,notonlyts7 butalsoTotolO101:ts7B5failedto

complementnine nsP2mutants

(five

with mutationspreviously

mapped to nsP2 and four subgroup II A

complementation

groupmutantswhose mutationswere

mapped only

recently

to

on November 9, 2019 by guest

http://jvi.asm.org/

(7)

nsP2[6a]) and tsl5,whose phenotypepredictsthat its mutation will also map tonsP2 (37). TotollOl:ts7B5 and ts4were not

defectiveinpolyprotein cleavage (data notshown) (37); thus, their failure toshow complementation with Agroupmutants wasnotaresult ofthetrappingofnonmutated nsPs to inactive precursors.Insummary,viruswithmutations in nsP4and nsPl complemented relatively efficiently viruses with mutations in nsP2and nsP3, butnone of 12viruses withmutations in nsP2 and nsP3 complemented each other (Table 3). Finally, Toto

1101:ts7B5 showed a difference from ts4 in its ability to

complement thensPlmutanttsll:the mutationatamino acid 268 in nsP3 ofts4blocked complementation,while the

muta-tionatamino acid 312 innsP3ofTotollOl:ts7B5didnot.Our interpretation of theseresults is that nsP3 associates with nsPl

andachange inaminoacid268of nsP3affectsthis association, while achange in amino acid 312of nsP3 doesnot. Such an association couldoccuratthelevelof the P123polyproteinor after the cleavage and release of nsPl. We favor the latter interpretation becauseofthe ability of the other nsP3mutants,

TotollOl:ts7B5 andts7, tocomplementtsll's defective nsPl.

DISCUSSION

Thisstudyshowed thatatslesion resultingfromachange of Ala-268toVal in the nsP3of SIN HRts4wasresponsiblefor the selective, rapid loss at 40°C ofnegative-strand synthesis, which was aphenotype that resembled the one possessed by thensPlmutanttsll of theBcomplementation groupofSIN

HR mutants. Therefore, both nsPl and nsP3 function in the synthesis of negative strands. The mapping of the mutation in a member of the A complementation group to nsP3 was unexpected, as the mutations in other A group mutants

mapped to nsP2 (16, 37, 39); however, it paralleled the

mapping of themutations inthetwoSIN HR G

complemen-tation groupmutants tsl8 and ts7 to nsP2 and nsP3,

respec-tively (16), and mayhave implications for theformation and structureof the earlyinitialreplication complex, asdiscussed

below.FindingthatnsP3functioned in negative-strand synthe-sis was surprising because there is no analog of nsP3 in the

plant viruses belonging to the SIN superfamily (19), which

would needto form replication complexes synthesizing

nega-tive strands. SIN nsP3 is a 549-amino-acid phosphorylated protein whose N-terminal 325 amino acids are highly con-served among the alphaviruses, while the C-terminal 225 aminoacidsarevariableorcanbedeleted (6,21,23, 32, 44,47, 48). The N terminus of nsP3 contains essential sequences, since deletion of amino acids 218 to273 is lethal (23, 26), as

wellassequencesthatarehighly conservedamong the

alpha-viruses, rubiviruses, hepatitis E virus, and coronaviruses and

flank their papain-like protease sequences (9, 12, 13). This

latter "X" domain, whichis located between amino acids 18

and 115 in SIN nsP3, has been suggested to function in polyprotein processing (1, 12, 13, 19). In the present study, aminoacid substitutions in nsP3at residues 268 and312 did notdrasticallyaffect overall processingof polyproteinsat40°C (datanot shown) (17, 20, 37) and therewas nomajor loss of

nsP4 functions that wouldhave been expected if polyprotein

processinghad been affected (24, 25).

One difference thatwe observedbetweentsll and ts4was

that the mutation in nsPl oftsll prevented the synthesis of

negative strands byastable replication complex,whereas the

mutation in nsP3 of ts4 did not. We conclude that nsPl

functioned in negative-strand synthesis by both the initial or

unstableformand thestable form of thereplication complex, whereas nsP3 functioned in negative-strand synthesis only by theinitial formof thereplication complex.This could resultif

the mutation is nsP3 of ts4 caused a ts lesion that wasexpressed in the precursor of nsP3 that would be part of the initial but not the stable replication complex. On the other hand, the muta-tion innsP1 of tsll expressed itself both in the precursor and in the stable, final form ofnsPl and affected both the initial and stable replication complexes. The temporal requirement for functional nsP3 would not reflect its transient physical presence because nsP3 was a component of active replication complexes late in infection (2). Thus, the effect of the ts lesion in nsP3 of ts4 on early negative-strand synthesis could be related to its effect on P23 (ts) versus nsP3 (no effect). Of 63 SIN HR ts mutants (3, 4, 46) isolated afterchemical mutagen-esis, only the mutations in ts4 and ts7 mapped to nsP3. Comparison of the two nsP3 mutants showed that the Phe-312 to Ser change in the nsP3 of ts7 caused negative-strand synthesis to cease more slowly than the Ala-268 to Val change inthe nsP3 of ts4. While ts7 failed to produce functional nsP3 or P23 proteins, it did produce nsPl proteins capable of complementing thensPl mutant tsll, while ts4 did not. This could reflect a greater leakiness of the Phe-312 to Ser lesion or could indicate that residue 312 is near but not within the functional or folding domain of nsP3 that would be affected by theAla-268 to Val mutation. Recently, two mutants with linker insertion mutations in the N-terminal 278 residues of nsP3 were found to resemble ts4 and ts7 in their RNA-negative phenotype, complementation pattern, and expression of a ts defect in negative-strand synthesis (22). Because they comple-mented tsll, we would predict that these mutants would resemble ts7, would cease negative-strand synthesis more slowly than tsll and ts4, and would allow reactivation of negative-strand synthesis.

Ints4, theobserved C-to-U transition is consistent with the original mutagenesis of the virus with nitrous acid (3). Al-though achange of Ala to Val might be considered minor since both arenonpolar (hydrophobic) amino acids and have similar isoelectric points(pI6.0), a change of Ala to Val in the SIN E2 glycoprotein led to defective virion assembly (15). Computer analysis (5, 10, 11) of the protein sequence predicted that the Ala-268 to Val change would convert this region from alpha to beta structure (Chou-Fasman and Garnier-Robson analysis), decrease its hydrophilicity (Kyte-Doolittle analysis), and lower its probability of being on the surface (Emini analysis). The Phe-312 to Ser change, on the other hand, was predicted to increase hydrophilicity and create a short coil region. Such structural alterations in the nsP3 of ts4 could account for the reduced levels of overall RNA synthesis and for the change in the ratio of 26S subgenomic RNA relative to 49S genome RNAsynthesis observed at the permissive temperature (20, 37), which arephenotypes that were not observed with ts7. The reduced level of overall viral RNA synthesis by ts4 at 30°C suggests that the region of nsP3 where the ts4 mutation is located isimportant for the formation of the initial replication complex. When synthesized and folded at 400C, the ts4 mu-tated nsP3 is completely inactive. Once the complex had formed at 30°C, however, the mutation did not affect positive-strand synthesis by the stable replication complex at 40°C or prevent stable replication complexes containing the nsP4 of ts24 from reactivating negative-strand synthesis. The unusual aspect of the ts4phenotype, compared with that of ts7, was the rapidity with which negative-strand synthesis ceased after exposure to high temperature. Several possibilities can be envisioned to explain these observations. The ts4 mutation could directly affect negative-strand synthesis by the initial replication complex or could act indirectly by altering the conformation and thus the activity of

nsPl.

In the latter case,

on November 9, 2019 by guest

http://jvi.asm.org/

(8)

I

nsPl

I

nsP2 X nsP3

I

nsP4

I

nsPlI nsP2 X | nsM

I

P1234, P123

initial

polyprotein

(X)

protease

P23

or

nsP2-nsP3 complex

]

nsP I

negative

strand

polymerase

xsP3

positive strand

polymerase

FIG. 6. Model for the processing of alphavirus polyproteins P1234 and P123 to form thenegative-strand polymerase and its subsequent

conversiontoapositive-strandpolymerase. The inabilitytodetectcomplementation between 12 nsP2 and nsP3mutantsarguesthesetwoproteins

mayfunctionasacis-active complextoformafunctional replication complex.Becausetheirexactfunctional form is unknown, theyarerepresented as aP23fusionprotein/polyproteinandas acis-active nsP2-nsP3 complex. TheXindicatesthepapain-likeproteasedomain in nsP2(7, 8, 13, 18).

nsP3would playarole in the foldingormodificationofother nsPs.

Examination of complementation by SIN RNA-negative mutantswas necessary to more fully assess the ability ofthe nsPstodissociate,exchange, andassemble with each other and

formreplication complexes.Complementation analysis ismore difficult when proteins areprocessed from apolyprotein and probably tend to associate with each other and not todiffuse

and interchange. For example, alphaviruses other than SIN

HRfail to show complementation and the SIN HRmutants

show only low levels of complementation. The polyprotein strategyfor theproduction of individual proteins thatforma complexisan efficient mechanismtoensurethat theproteins

associate. Plant viruses belonging to the Sindbis superfamily produce homologous nsPs that actually function as polypro-teins. In spite of thesedifficulties, the results ofour comple-mentation analysisconfirmed thereportby Hahn et al. (16),

whofoundthatts7 (originally placedin theGgroup) didnot complementtsl8(originally defined as aG groupmutantbut withitscausal mutation mappedtonsP2)orts24(anAgroup

mutantwithadefectivensP2).Furthermore,wecompletedthe

analysisand found that this failurewas aconsistent property of allBurgeandPfefferkorn (3, 4) andStrauss and Strauss (46) nsP2 and nsP3mutants.No nsP3mutantscomplemented any

ofthe nine nsP2mutants,whichhavemutationsmappingover atleast60% of the nsP2protein,ortslS,whichis a presump-tive nsP2mutant. This resultarguesthatnsP4 and nsPlwere diffusable and activeintrans and thattheircleavagefrom the nascent P1234 and P123 polyproteins occurred early in the

processofformingtheinitialreplication complex.Ontheother hand,nsP2 and nsP3 wouldactas asinglecistronbecausethey

formed a stable association immediately after their synthesis

andpresumablybefore theircleavageorbecause the

polypro-tein P23was the functional form. Therefore, although there are four viral nonstructural proteins, there are actually only threecomplementationgroupsfortheoriginalSIN HR RNA-negative mutants. The only exception to this would be the

recent report oftwo SINlinker insertionmutantscarrying an extrasixamino acids between residues 58 and 59 or226 and 227of nsP3 (22). The firstmutantcomplemented ts18, which

hadamutationincodon 509 for nsP2, butnottsl7,which had a mutation in codon 517 for nsP2, while the latter mutant complemented both the nsP2 mutants but to a much lower

extent than the nsPl mutant tsll and the nsP4 mutant ts6.

Unlikesingle-residue substitutions,theadditionofaminoacid

residuesmaydistort and slowthefolding of thenonstructural proteinsandallowincreasedopportunityforcomplementation tooccur.Thiswouldexplain thehigh levels of

complementa-tion observed with these mutants. It would also argue that whileasinglecistron form likeP23maybefunctional,it isnot

required for the activity of replication complexes; nsP2 and nsP3 canalsoprovide activity.

nsP4,not P34,isthefunctionalform of theGDD replicase

sequence (2, 24, 39),and its release from the P1234

polypro-tein is necessary for transcriptional activation (24, 25, 45).

Recentanalysesofcleavage-defective variants ofSIN demon-strated that fully cleaved nsPl, nsP2, nsP3, and nsP4 form

replication complexesthatareactive inpositive-strand synthe-sis (25, 45),while polyprotein P123 and nsP4were active in negative-strand synthesis.These results supportour discovery that replication complexes found late in infection contained fullyprocessednsPs (2). From the resultspresentedhere,we suggest that the initial replication complex that possesses negative-strand polymerase activityiscomposedofnsP4, nsPl,

and P23 or tightly associated nsP2 and nsP3 (Fig. 6). The

ability oftsll, which has a mutated nsPl and is defective in

negative-strand synthesis, to complement all other mutants exceptts4stronglyarguesthatcleavednsPlcanassembleinto

replication complexesthatengageinnegative-strand synthesis.

Thus, cleavage of nsPlfrom thepolyprotein normallywould be anearlyevent. Moreover,arequirementforpolyproteinP123 or P12 (17) is unlikely because cleaved nsPl functions in

negative-strand synthesis late in infection (35, 50). Thus, the

current evidence suggests that, depending on conditions,

nsPl +/-hostfactors

nsPi

h If

C

+/-host factors

nsP2

x

Fr..

on November 9, 2019 by guest

http://jvi.asm.org/

(9)

cleaved or uncleaved forms of nsPl, nsP2, and nsP3 plus cleavednsP4canassumethecorrect conformation andinteract to function in negative-strand synthesis. Once formed, the short-lived nature of the negative-strand polymerase activity results from theconversion of the initialreplication complexto a stable one that synthesizes positive strands by using the negative strand as a stable template (36, 38, 40). Conversion

from the initial tothe stable complex could be accomplished viacleavageof P23and/orrearrangement ofannsP2plus nsP3 complex. Others have proposed recently that a cleavage of P123 converts the replication complex into one possessing

positive-strand polymerase activity (25, 45). Domains located to nsP2 and nsP4 (35, 39) affect the level of promoter

recognition by the stable replication complex and determine the synthesis ofpositive strands relative to negative strands. Thus, the interaction of nsP2 and nsP4 may be affected by conformational changes that would result from mutations in

nsP2 or nsP3. Although it had been suggested that the

temporal synthesis ofnegative strands resulted from a

regu-latedprocessingofthensPpolyproteins (17, 44), inparticular

from a change in the cleavage pattern of P1234 to form P34 andnot nsP4late ininfection (7), muchevidence nowargues

against this model (24, 25, 39, 45). Why negative-strand synthesis ceases early in infection is still unanswered. It may cease because ofthe failure to form P23 or P123, because of their more rapid cleavage late in infection, or because the exhaustionofahostfactor(s) (2, 29-31) limitstheformationof initial replication complexes late ininfection.

Mutations ineach ofthefournsPshavebeenshowntogive an RNA-negative phenotype. Although all alphavirus RNA-negative mutantsfail toproduce stable replication complexes thatareactiveinpositive-strand synthesis at40°C, they donot share the distinct phenotype of ts4 and tsll, which is an immediate effect on negative-strand synthesis after a shift to the nonpermissive temperature. Therefore, other alphavirus RNA-negativetsmutantsthatarenotdefectivein cleavageof the P1234 polyprotein and do not show as rapid a loss of negative-strand synthesisasts4 andtsllshould showdefects in theformation ofreplication complexes thatoccur afterthose caused by the ts4 and tsll mutations. Such ts lesions would preventtheswitchfromnegative-topositive-strandsynthesis,

i.e.,theconversionofaninitialtoastablereplicationcomplex. Theabilityofstablereplication complexescontainingnsP4 of ts24 to reactivate negative-strand synthesis has allowedus to discriminate between ts lesions that affect negative-strand synthesis specifically by the initial replication complex (ts4)

from those that affect overall negative-strand polymerase

activity(tsll).Thishasproventobeapowerful tooland will be especially useful in identifying mutants that are defective specifically in the formation of initial replication complexes, which requires negative-strand synthesis,orin the conversion ofinitialtostable replication complexes.

ACKNOWLEDGMENTS

Wegladly acknowledgethetechnical assistanceof RebeccaCastle andthe giftofTotollOl:ts7B5 cDNA from E. G. Strauss andJ. H. Strauss.

Supportfor these studieswas derived fromPublic Health Service

grantAI-15123 fromtheNational Institutesof Health. REFERENCES

1. Ahlquist, P.,E.G.Strauss, C.M. Rice, J. H. Strauss,J.Haseloff,

and D. Zimmern. 1985. Sindbis virus proteins nsPl and nsP2 contain homology to nonstructural proteins from several RNA

plantviruses.J.Virol.53:536-542.

2. Barton,D.J.,S.G.Sawicki, and D. L. Sawicki.1991. Solubilization

and immunoprecipitation ofalphavirus replicationcomplexes. J.

Virol. 65:1496-1506.

3. Burge, B.W., and E.R. Pfefferkorn. 1966. Isolation and

charac-terization of conditional-lethal mutantsofSindbis virus.

Virology

30:204-213.

4. Burge, B. W., and E. R. Pfefferkorn. 1966. Complementation

betweentemperature-sensitive mutantsofSindbisvirus.

Virology

30:214-223.

5. Chou,P.Y.,and G. D. Fasman.1978.Prediction of thesecondary

structure of proteins from their amino acid sequence. Adv.

Enzymol. Relat. AreasMol. Biol. 47:45-148.

6. Davis, N. L., L. V. Willis, J.F.Smith,andR.E.Johnston. 1989. In

vitro synthesis of infectious Venezuelanequineencephalitisvirus RNA from a cDNA clone: analysis ofa viable deletion mutant. Virology 171:189-204.

6a.De, I., S. G. Sawicki,and D. L. Sawicki. Unpublisheddata.

7. deGroot, R. J., W. R. Hardy,Y.Shirako,andJ.H.Strauss. 1990.

Cleavage-site preferences of Sindbis viruspolyproteinscontaining

thenon-structuralproteinase: evidence fortemporal regulationof

polyprotein processingin vivo. EMBO J.9:2631-2638.

8. Ding, M.,and M.J. Schlesinger. 1989. Evidence that Sindbis virus nsP2is an autoprotease which processes the virus nonstructural polyprotein. Virology 171:280-284.

9. Dominguez, G.,C.-Y.Wang,and T. K.Frey.1990.Sequence of the genomeRNAofrubellavirus:evidence forgeneticrearrangement during togavirus evolution. Virology 177:225-238.

10. Feng, D. S., M. S. Johnson, and R. F. Doolittle. 1985. Aligning

amino acidsequences: comparisonofcommonlyusedmethods.J.

Mol. Evol.21:112-125.

11. Garnier, J.,D.J. Osguthorpe,and B. Robson.1978.Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. Mol. Biol. 120:97-120. 12. Gorbalenya, A.E., and E. V. Koonin. 1993. Comparative analysis of amino-acid sequences of key enzymes of replication and expression of positive-strand RNA viruses: validity ofapproach and functional and evolutionary implications. Soc. Sci. Rev. D Physicochem. Biol.11:1-84.

13. Gorbalenya,A.E., E.V. Koonin,and M. M.-C. Lai. 1991. Putative papain-related thiol proteases of positive-strand RNA viruses. FEBSLett. 288:201-205.

14. Grady, L. J.,andW. P. Campbell. 1989. Amplification of large

RNAs (>1.5 kb) by polymerase chain reaction. BioTechniques 7:798-800.

15. Hahn, C.S.,C.M. Rice, E. G. Strauss, E. M.Lenches,and J. H. Strauss. 1989. Sindbis virustslO3 has a mutation in glycoprotein

E2thatleadstodefective assembly of virions. J. Virol. 63:3459-3465.

16. Hahn,Y.S., E.G. Strauss,and J. H. Strauss. 1989. Mapping of

RNA-temperature-sensitivemutants of Sindbis virus: assignment

ofcomplementationgroupsA, B,andG to nonstructural proteins. J.Virol.63:3142-3150.

17. Hardy,W.R.,Y. S.Hahn,R.J. de Groot, E. Strauss, and J. H.

Strauss. 1990. Synthesis and processing of the nonstructural polyproteinsofseveral temperature-sensitive mutants of Sindbis

virus.Virology177:199-208.

18. Hardy, W.R,and J.H.Strauss. 1989.Processing the

nonstruc-turalpolyproteins ofSindbisvirus:thenonstructuralproteinase is

in theC-terminal half ofnsP2 and functions both in cis and in

trans. J.Virol.63:4653-4664.

19. Haselof, J., P. Goelet, D. Zimmern, P. Ahlquist, R Dasgupta, and P. Kaesberg. 1984. Strikingsimilarities in amino acid sequence

amongnonstructuralproteinsencodedby RNA viruses that have

dissimilar genomic organization. Proc. Natl. Acad. Sci. USA

81:4358-4362.

20. Keranen, S., and L. Kiiariainen. 1979. Functional defects of

RNA-negativetemperature-sensitivemutantsof Sindbis and

Sem-likiForest viruses. J. Virol.32:19-29.

21. Kinney, R M., B. J. B. Johnson, J. B. Welch,K.R Tsuchiya, and D. W. Trent. 1989. The full-lengthnucleotide sequences of the

virulent Trinidad donkeystrain ofVenezuelanequineencephalitis

virus and itsattenuated vaccinederivative,strainTC-83. Virology

170:19-30.

22. LaStarza,M.,J. Lemm,andC.M. Rice. 1994. Genetic analysis of

on November 9, 2019 by guest

http://jvi.asm.org/

(10)

the nsP3region ofSindbisvirus: evidencefor roles in minus-strand andsubgenomic RNA synthesis. J. Virol. 68:5781-5791. 23. LaStarza, M., and C. M. Rice. 1994. Deletion and duplication

mutations in the C-terminalnonconserved region of Sindbis virus nsP3: effects on phosphorylation and on virus replication in vertebrate and invertebrate cells. Virology 202:224-232. 24. Lemm, J. A., and C. M. Rice. 1993. Roles of nonstructural

polyproteins and cleavage products in regulating Sindbis virus RNAreplication and transcription. J. Virol. 67:1916-1926. 25. Lemm,J.A., and C. M. Rice. 1994. Polypeptide requirements for

assembly of functional Sindbis virus replication complexes: a model for thetemporal regulationofminus and plusstrand RNA

synthesis.EMBO J. 13:2925-2934.

26. Li, G., M. W. LaStarza, W. R. Hardy, J. H. Strauss, and C. M. Rice. 1990. Phosphorylation of Sindbisvirus nsP3 in vivo and in vitro.Virology 179:416-427.

27. Lopez, S., J. R. Bell, E. G.Strauss, and J. H. Strauss. 1985. The nonstructuralproteins of Sindbis virus asstudied withanantibody

specific for the C terminus of the nonstructural readthrough polyprotein. Virology141:235-247.

28. Mi, S., and V. Stollar. 1991.Expression of SindbisvirusnsPland methyltransferase activity in Escherichia coli. Virology 184:423-427.

29. Nakhasi, H. L., X.-Q. Cao, T. A. Rouault, and T.-Y. Liu. 1991.

Specific bindingofhostcellproteinstothe3'-terminal stem-loop

structureof rubellavirus negative-strandRNA. J.Virol. 65:5961-5967.

30. Nakhasi, H. L., T. A. Rouault, D.J. Haile, T. Y. Liu, and R. D. Klausner. 1990.Specifichigh-affinity binding of hostcellproteins

tothe3'region of rubella virusRNA.NewBiol.2:255-264. 31. Pardigon,N., and J. H. Strauss. 1992.Cellularproteins bind to the

3' end ofSindbisvirusminus-strandRNA. J. Virol. 66:1007-1015. 32. Peranen, J. 1991. Localization and phosphorylation of Semliki Forest virus non-structural protein nsP3expressed in COS cells

from aclonedcDNA. J. Gen. Virol. 72:195-199.

33. Rice, C. M., R. Levis, J. H. Strauss, and H. V. Huang. 1987. Production of infectious RNA transcripts from Sindbis virus cDNAclones: mappingoflethalmutations, rescue of a

tempera-ture-sensitive marker, and invitro mutagenesis to generate

de-finedmutants.J. Virol. 61:3809-3819.

34. Sanger, F.,S.Nicklen,and A. R.Coulson. 1977. DNAsequencing

with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA

74:5463-5467.

35. Sawicki,D.L.,D. B.Barkhimer,S. G.Sawicki,C. M.Rice,andS.

Schlesinger. 1990. Temperature sensitive shut-off of alphavirus

minus strand RNA synthesis maps to a nonstructural protein,

nsP4.Virology 174:43-52.

36. Sawicki,D.L., and S. G. Sawicki. 1980. Short-lived minus-strand

polymerase for Semliki Forest virus. J. Virol. 34:108-118. 37. Sawicki,D.L., and S. G. Sawicki. 1985.Functionalanalysisof the

Acomplementation group mutants of Sindbis HR virus.Virology

144:20-34.

38. Sawicki, D. L., and S. G. Sawicki. 1987. Alphavirus plus and minus strandRNAsynthesis, p.251-259. In M. Brinton and R. Rueckert

(ed.), Positive-strandRNAviruses. Alan R.Liss, Inc., New York. 39. Sawicki,D. L., and S. G. Sawicki. 1993. A secondnonstructural protein functions in the regulation of alphavirus negative-strand RNAsynthesis. J. Virol.67:3605-3610.

40. Sawicki,D.L., S. G. Sawicki, S. Keranen, and L. Kaariainen. 1981.

SpecificSindbisvirus-coded function forminus-strand RNA

syn-thesis.J.Virol.39:348-358.

41. Sawicki, S. G., and D. L. Sawicki. 1986. Theeffectof overproduc-tion of nonstructuralproteinsonalphavirus plusstrand andminus strand RNAsynthesis. Virology152:507-512.

42. Sawicki, S. G., D. L. Sawicki, L.Kiiariainen,and S.Keranen.1981. ASindbisvirus mutanttemperature-sensitiveintheregulationof minus-strand RNAsynthesis. Virology115:161-172.

43. Shirako, Y., B. Niklasson, J. M.Dalrynple,E.Strauss, and J. H. Strauss. 1991. Structure of the Ockelbo virus genome and its relationshiptootherSindbis viruses. Virology182:753-764. 44. Shirako, Y., and J. H. Strauss. 1990.CleavagebetweennsPl and

nsP2 initiates the processingpathway of Sindbis virus nonstruc-turalpolyprotein P123.Virology177:54-64.

45. Shirako,Y., and J. H. Strauss. 1994.Regulationof Sindbis virus RNAreplication: uncleaved P123 and nsP4 function in

minus-strandRNA synthesis, whereascleavedproducts from P123 are

requiredforefficient plus-strandRNAsynthesis.J.Virol. 68:1874-1885.

46. Strauss,E.G.,E. M.Lenches, andJ. H. Strauss. 1976. Mutantsof Sindbis virus. I. Isolationand partial characterizationof 89new temperature-sensitivemutants.Virology74:154-168.

47. Strauss,E.G., R.Levinson, C. M.Rice, J.Dalrymple,andJ.H. Strauss. 1988.NonstructuralproteinsnsP3 and nsP4of Ross River

and O'Nyong-nyong viruses:sequence andcomparisonwith those

ofotheralphaviruses. Virology164:265-274.

48. Strauss, E. G., C. M. Rice,and J. H. Strauss. 1984. Complete

nucleotide sequence of thegenomicRNAofSindbisvirus. Virol-ogy 133:92-110.

49. Wang, Y.-F. 1992.Identificationandanalysisofproteins essential foralphavirus negative-strand RNAsynthesis. Ph.D. thesis.

Med-icalCollegeofOhio,Toledo.

50. Wang,Y.-F., S. G.Sawicki,and D. L.Sawicki. 1991. Sindbisvirus nsPlfunctions innegative-strandRNAsynthesis.J.Virol. 65:985-988.

51. Weiss, B. G., and S.Schlesinger. 1991. Recombination between

Sindbisvirus RNAs. J. Virol. 65:4017-4025.