Sequence Determination of the (+) Leader RNA Regions of the Vesicular Stomatitis Virus Chandipura, Cocal, and Piry Serotype Genomes

0022-538X/83/040125-06$02.00/0

Copyright01983,AmericanSocietyforMicrobiology

Sequence Determination of the (+) Leader RNA Regions of

the

Vesicular Stomatitis Virus Chandipura, Cocal, and Piry

Serotype Genomes

COLOMBAGIORGI,tBENJAMIN BLUMBERG, AND DANIEL KOLAKOFSKY* DepartmentofMicrobiology, University of Geneva Medical School, 1205 Geneva, Switzerland

Received 30 November 1982/Accepted 12 January 1983

Using 3'-end-labeled genome probes, cells infected with vesicular stomatitis virus Chandipura, Cocal, and Piry serotypes were shown to contain (+) leader

RNAsof approximately 50 nucleotides in length.Thenucleotidesequenceofthe

leader RNA regions ofthese genomes was determined and compared with the

previously reported sequences of both the (+) and (-) leader RNA regions of

other vesicular stomatitis virus serotypes. Regions ofstrong conservation of nucleotide sequence among the various vesicular stomatitis virus serotypes suggest those nucleotides thought to be involved in control functions during

vesicular stomatitis virus replication.

Vesicular stomatitis virus (VSV), a rhabdo-virus,

contains

all

of

its

genetic information

(11

to 12

kilobases)

ina

single

RNAchain of

nega-tive

polarity

(-). The VSV (-) genome codes for five viral

proteins

which are

expressed

via fivemonocistronic mRNAs. These mRNAs are

transcribed in the order 3' N, NS, M,

G,

andL

5' both in vivo and in vitro, and the relative molar abundanceof each mRNA is determined by its

proximity

to the 3' end of the genome,

which is thoughttobe theonly siteatwhich the viral polymerasecan enterthe

genofpe temnplate

(1, 21). The first RNA transcribed from the (-)

genome

template

is not the

first

mRNA but a

short,

47-base

(+)

leader

RNA,

which is

comple-mentary tothe

precise

3' endof the (-)genome

(4).This leader RNA hasrecently been shownto

contain the nucleation site for the initiation of

nucleocapsid

(NC) assembly

(la-3). One

possi-ble function

of

the (+) leader RNA is thus to

detachthis nucleation site from the

body

of the first mRNA so thatthis mRNAwould notalso be

encapsidated.

A model for VSV

replication

has

recently

been advanced in which the

binding

ofN

protein

to the nascent leader RNA chain modulates

mRNAtranscription andgenomereplication (3, 11). In this model, the termination site forthe

leaderRNAis

analogous

totheattenuatorof the tryptophanoperonof bacteriaand the

binding

of

Nproteintothe nascentleaderchain would both

promote readthrough of the leader termination

signal

and

simultaneously begin

the

encapsida-tion of the nascent

antigenome

strand. Thus,

tOn leave fromLaboratoriodiVirologia,IstitutoSuperiore diSanita,Rome, Italy.

manyof the features which control VSV

replica-tion

arethoughttobe encoded in the relatively short leader RNA

region.

In aneffort tobetter understand how these control processes might

operate, wehave

sequenced

the leader regions of the

Piry

(Pi),

Chandipura (Ch),

andCocal(Co)

serotypes of VSV.

MATERIALS AND METHODS

Viruses and theirgrowth. VSV Indiana(In; Mudd-Summers strain)was obtained from Donald F. Surn-mers, University of Utah School ofMedicine, Salt LakeCity. VSVCh and Cowereobtained fromCraig Pringle,Institute of Virology,Glasgow,Scotland,and VSV Pi was obtainedfrom DanielleTeninges, Labora-toire Genetique des Virus, Gif-sur-Yvette, France. Each serotype was plaque purified twice in

succes-sion,and asingle plaque was thenused to generate virusstocks on five 10-cm dishes of BHK cells. Virus

wasgrown oneither BHK orLLCMK2 cells (10 PFU/

cell) and harvestedat18 hpostinfectionfor all

sero-types exceptCo (36 hpostinfection).Radioactive virus wasgrownin the presenceof 20 ,iCi of[3H]uridineper ml and 0.5

"ag

ofactinomycinD per ml. The viruswas

thenpelleted through30%oglycerol, suspended in TNE

buffer,banded onaglycerol-tartrate gradient(16), and then recoveredbypelleting.Asodiumdodecyl

sulfate-polyacrylamidegelelectrophoresispatternoftheviral

proteinsof each serotype is shown inFig. 1.

Preparation of3'-end-labeled virion RNA (genome probe).Purified virus (1 to 3mg/ml) was incubated for 20minat37°C with0.5% sodium dodecyl sulfate and 500 ,ug of proteinase K per ml and then directly

centrifuged on 5 to23% sucrosegradients (15). The 42S RNA was localized by counting samples of the gradient for13H]RNA andrecovered by two ethanol precipitations. Virion 42S RNA (5 to 10

Rg)

was then 3' endlabeled with32P-labeledpCp andRNAligase in

a 20-j, reaction as previously described (15) and 125

on November 10, 2019 by guest

http://jvi.asm.org/

chromatographed on Sephadex G-50, and the excluded RNA was ethanol precipitated. The precipitate was dissolved in 25 ,ul of 10 mM Na2PO4, denatured with 40 mM CH3HgOH, and then diluted to 100 ,ul with TNEcontaining 100 mM BME and recentrifuged on a sodium dodecyl sulfate sucrose gradient. RNA which sedimented at 42S was again selected and recovered by ethanol precipitation.

Sequence determination. (i) Purine reactions. Partial digestion with ribonucleasesT1 and U2 was carried out inbuffercontaining 20 mM sodium citrate (pH 5.0 for

T1,pH 3.5for U2), 7 M urea,1mMEDTA, 0.25 mgof tRNA perml, 0.25% xylene cyanol,andbromphenol

blue (5,12). Theend-labeled RNAwasaddedtothis buffer, and thereaction mixture wasdistributed into three 20-,ul portions in siliconized Eppendorftubes. The tubes were heated at50°C for 5 min and quick

chilled in ice. A 1-,ulportionof ribonucleaseT1(10-1 U/ml)orU2(10-2U/ml)wasaddedtothefirst of the three tubes, and a single 10-fold serial dilution was

carriedoutinto the second tube. The third tubewas

incubated without enzyme. The tubeswereincubated for 15 minat50°C,chilled,and loadedonto12or20%

sequencinggels.

Fortheladder, 1.5 ,ul of end-labeled RNAplus2,ul of tRNA(2,ug/ml)wasmixed with 15 ,ul offormamide. Thereactionwasincubatedat100°Cfor 10to25 min.

(ii) Pyrimidinereactions. Forthe Ureaction,the 3' labeled RNAwaslyophilizedinthe presence of tRNA

carrier, dissolved in 50%ohydrazine-50%o water(vol/

vol), andincubated on ice for 10to15min(17). TheC> Ureaction wascarried outin 3 M NaCl-hydrazine incubatedonice for 20to30min(17).The

chemicallymodified RNAwasprecipitatedtwice with ethanol in the presenceof0.3Msodiumacetate.Itwas

thendissolved in 20 ,ulof 1.0 M aniline-acetatebuffer,

pH4.5, and incubated in the darkat60°Cfor 20 min. The reaction was terminated by freezing at -70°C.

The samples were lyophilized, dissolved in water, chilled at -70°C, and lyophilized again. Finally, the

samplesweredissolved in 3to5,ul of 8 M urea-20 mM

Tris-hydrochloride (pH 7.4)-i mMEDTA-0.05% xy-lene cyanol and bromphenol blue, heated 2 min at

90°C,and loadedontothesequencing gel.

Co

Pi

Ch

In

L

uAIb

N

FIG. 1. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the various VSV serotypes. Be-tween10to20 p.gof thevarious VSVserotypeswere

electrophoresedona 10oacrylamide-0.8% bis(acryl-amide) protein gel (13) whichwas then stained with

Coomassie brilliant blue. Theserotypesare notedat thetopof the gel. The positions of theL,G,N,and M proteinsof VSV Inareindicatedonthe right.

RESULTS

VSV Ch, Co, and Pi also contain (+) leader RNAs. To date, only the In and New Jersey serotypes of VSV have been shown toinitiate transcriptionwith a (+) leader RNA. To investi-gate whether the Ch, Co, and Pi serotypes

similarly contained leader RNAs, weprepared 42S (-)RNAfrom each of the various purified virions (Fig. 1), and these (-) genomes were

radiolabeledattheir 3'OHends with

32P-labeled

pCp andRNAligase (see above). Thepresence of(+) leader RNA in virus-infected cells was examined by annealing CsCl pellet RNA (total

cytoplasmic RNA from which viral NCs have been removedby CsCl density gradient centrifu-gation [15]) against the end-labeled genome

probes,

followed

by ribonuclease digestion to

degradesingle-stranded regions ofthe RNA. As a control, a similar optical density amount of

CsCl

pellet

RNA from

uninfected

cells was

annealed in

parallel.

The

polyacrylamide

gel of the

remaining

RNase-resistant double-stranded

RNAsfrom thisexperiment is shown in Fig. 2. Note that when uninfected CsCl pelletRNA is

annealedagainst the various 3' genomeprobes, only in thecaseof the Cogenome isanyof the labeledgenome RNA resistant to RNase

diges-tion. Thisshortdouble-stranded RNAprobably

represents the complementary ends of the Co

genomeitself, similarto that seenin VSVNew

Jersey (7). However, when CsCl pellet RNAs from the various virus-infected cells were an-nealedagainst their homologousgenomeprobes,

severaldouble-stranded RNAs resulted in each

case.Theresults with VSVCh andInappeared identical; three bands of decreasing intensity which migrated just in advance ofthe 75-base

4b

on November 10, 2019 by guest

http://jvi.asm.org/

[image:2.490.279.425.73.393.2]

SEQUENCE

OF THE VSV (+) LEADER RNA 127

A C D E F G H J K

--BPB

FIG. 2. Demonstrationof ()leader RNAs in the

various VSV serotype-infected cells. CsCl pellet

RNAs from the various VSV-infected BHKcells or

from uninfected cells were annealed against their

homologous 3'-end-labeled genomeprobes. The

reac-tionswerethendigestedwith 20 p.gof RNase A per ml

in 0.375 MNaCi aspreviouslydescribed(15),and the

remaining RNAwas recovered by ethanol

precipita-tion andelectrophoresed on a acrylamide-0.5%

bis(acrylamide) gel (0.4mmby20cmby40cm).Lane

A, Hinfdigest ofpBr322; lanes B to D, 0.5 and 1.5

optical density units (ODU) of VSV In-infected and

0.5 ODU of uninfected CsCl pellet RNA annealed

againstthe VSV In genomeprobe; lanes Eto G,0.5

and 1.5 ODU of VSV Co-infected and 0.5 ODU of

uninfected CsClpelletRNAannealedagainstthe VSV

Co genomeprobe;lanes HtoI, 1.25ODU of VSV

Pi-infected and 1.5 ODU of unPi-infected CsClpelletRNA

annealedagainstthe VSV Pi genomeprobe;lanesJto

K, 1.5 ODU of VSV Ch-infected and 1.5 ODU of

uninfected CsClpelletRNAannealedagainsttheVSV

Ch genome probe. XC, Xylene cyanol; BPB,

brom-phenolblue.

pair DNA marker resulted from the annealing. These bands were in the expected position of

leader double-strandedRNA onthesegels (10). However,the presence of three double-stranded RNAbands instead ofasingleband is probably

duetothehigher resolution of the long and thin

gels used in this study. To more accurately determine the chain length of the 3' genome

RNA protected against RNase digestion by

these leader RNAs, the RNA was recovered from each of the three In and Ch bands, and a

sample of these RNAs was partially digested with RNase T1. The partial digests were then

electrophoresed along with untreated band

ma-terial andaformamide ladder of the homologous end-labeled genome probes on a7 M urea

se-quencing gel (data not shown). From the posi-tion of the guanine nucleotides in the sequence

relative to the ladder, we estimate that the lengths of the 3'-end-labeled genome RNAs were42 + 43bases for the bottom band,44+45

bases for the middleband, and 47+51bases for thetopband(bases 48to51 ineachsequence are

purines and cannot be cut by RNase A; see

below).

The

remarkably

similar pattern between VSV In and Ch shown inFig. 2parallelsavery

similarsequence atthe 3' end of theirgenomes;

only 2outof the first 51 nucleotidesaredifferent between these two serotypes (see below). A

similar length

heterogeneity

of theIn(+)leader

RNA has recently been demonstrated directly by Schubert et al. (19)

by

directly labeling in

vitro transcripts with [3- 2P]ATP. Thus, VSV Ch-infected cells, like VSV In-infected cells, contain (+) leader RNAs which are 42 to 47

bases inlength.

The annealing of both Pi and Co CsCl pellet

RNA against their homologous genomeprobes also resulted in multiple double-stranded RNA bands. In the case of VSV Co, two major double-stranded RNA bandsare seeninaddition

to the putative complementary ends ofthe Co

genomeitself; thetopbandmigrates just slightly slower thanthe topbandof theInandChseries, and the faster band inaseparateexperimentwas

estimated to be

approximately

26 base

pairs

long. The reason why the shortest band

in-creases in

intensity

on

annealing with

infected cell RNAisnot

clear,

but itmaybe dueto athird

leader RNA of this size. In the case of VSVPi, atleast nine bandscouldroutinely bedetected, of which the topband migrated similarlytothe topbandofthe In andCh series. VSVCo-and

VSV Pi-infected cells therefore also contain

leaderRNAs,but these leader RNAs are

appar-ently heterogeneous over a wider

length

than their In andChcounterparts. It should benoted, however,thatalthough theinfectingstocks used

in these experiments were prepared from plaque-purified virus, the presence of small

VOL.46,1983

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.490.47.240.94.445.2]

A B C D E F G H J

... ... " ,

..

-A.

so

0 -EUXC

-

-l0.

[image:4.490.67.233.76.465.2]

--...9,N

...

FIG. 3. Sequence determination of3'-end-labeled

VSV Pi genome.The conditions for partialdigestion at specific bases are described in the text. Lane A,

formamideladder (10 minat100°C);lane B, chemical reaction for U; lane C, chemical reaction for C> U; laneD, formamide ladder (25 min at 100°C); lane E, 10- Uof RNase U2; lane F,10-2Uof RNase U2; lane

G, formamideladder (10 minat100°C), lane H,10' U of RNase

T1;

lane I,10-2Uof RNase

Tj;

lane J, no enzymeadded. BPB and XC on the right-hand margin show thepositions of the bromphenol blue and xylene cyanol dye markers, respectively. Dots on the right-hand side of given bandsindicate those bands which determine the sequence. The positions of the chemical

cleavageproducts containing 5' phosphates relativeto

theladder bands containing5'OH groups were deter-mined by sequencing VSV In on a separate gel. The dots show the sequence from the G at position 2 (lowest band, lanes H andI)tothe AAAAsequence at

positions 46 to 49, except for the U residues at

positions39, 40, 42, 43, and44whicharedetectableon

the film but not on the photograph of this

20%o

amountsof defective interferinggenomes

gener-ated inourinfections cannot be excluded. It is thereforepossiblethat some of the smaller dou-ble-stranded RNA bands detected in the Pi and Coexperimentswereduetoincomplete

anneal-ingwith (-)leader RNA.

Thus,cellsinfected with each of various VSV serotypes examined contained leader RNAs

which were complementary to the precise 3' ends oftheir(-)genometemplates. Thelargest

leader RNA in each case wasapproximately 50 baseslong.Inaddition,results identical to those shown inFig. 2 were obtained by annealing the products of in vitro virion transcription reac-tions to their homologous genome probes (not

shown).

Sequence determination of the (+) leader

re-gionsofVSVCh, Co, andPi. To date, only the

(+) leader regions

of

the In and New Jersey

serotypesof VSV have been sequenced. It was therefore of interest to determinethenucleotide sequence at the 3' ends ofthe Ch, Co, and Pi genomes,sinceannealing experiments have pre-viously shown that there is onlylimited homolo-gy between the coding regions of the various VSVserotypes (18).

The sequence determination was carried out on the 3'-end-labeled (-) genomes, using RNasesT1 andU2 to determine the position of G andArespectively, chemicalcleavageto deter-mine C andU, andformamide laddersto deter-mine chainlength (see above). Toposition the chemical cleavage products which contain a 5' phosphate relative to the ladder bands which have a 5' OH, the known VSV In genome was sequenced in parallel. The 3' terminal base, which was difficult to read from the sequence gel, was determinedby nearestneighbortransfer after alkalinehydrolysisand foundtobe U in all cases. Anexample of the sequencing gels, that for VSVPi, is shown inFig. 3. Thesequence of the 3' ends of VSVCo, Ch,and Piareshown in

Fig.4alongwithpreviously reported sequences

of VSV In and VSV NewJersey, and also the (-) leader regions of three In variants [which represent the 3' region of the (+) antigenome

strand].

DISCUSSION

The nucleotidesequences atthe 3' end of the

(-)genome of five VSVserotypes and of the 3' end ofthe (+)

antigenome

of three variants of the Inserotypehave severalinterestingfeatures.

Of theeightsequencesshown,allof whichdiffer

from eachother, there isremarkable homology

acrylamidesequencinggel. The various VSV genome sequences were alsodetermined on12% acrylamide gelstobetter resolve the bands above40nucleotides in length.

on November 10, 2019 by guest

http://jvi.asm.org/

SEQUENCE OF THE VSV (+) LEADER RNA

VSV GENOME RNA SEQUENCES

3' HO-UGCUICGUlUGUUUAAUMAAUAGUAAUUUUCCGAGUCCUCUUUGAAA

3' HO CUGUUUUGAAUGU UAUAACCGGAUCUCCCUUGAAA

3'

HO-UGCUUCUUIUGUW

UAAUUUGUC MUAUAUCAUCAUGUGC

3' HO-UGiCUCWIJUUGUUUGGUUAUCGUAAGAAUAUUCUGUAUUCUUUGMAAM

3' HO-UGCUUGU1UGUUAUAUAUUUUUCCGAGUCCUCUUUGAAA

3'

HO-UlUUCUGGUGUUUlGUCUAUUUUUUAUUVY,GYGUUCUCCCAG

3' HO-U

UUCUGGUGUUUUGMCUGUUUUUUAUUUVV,V,GUUCUCCCAG

3'

HO-UfGC

UGUUGUUUUGGUCGAUUUUUUGUUUUUGGUGUCUUCCCAG...$les

129.

IN MS

-NJ

-COCAL

-PIRY

-CHANDIPURA

-IN MS+

IN TSG 31 +

IN CAR4 +

FIG. 4. Nucleotide sequences at the 3' ends of various VSV genomes and antigenomes. The various serotypesarenotedatthe right-handside, where (-) and (+) denotegenomesandantigenomes, respectively.

The In(-),NJ(-), In(+), In tsG31(+),andInCAR4(+)sequences arefrom references(14,20). Theregions of

sequenceconservationatboth ends of the leader regionsareunderlined. The repetition of the hexanucleotide

UUUGGUisdottedunderneath.

within the first 18 nucleotides and then the sequencesdiverge widely. The regions of abso-lutehomology within the first 18 nucleotidesare underlined inFig. 4 and give risetothe

consen-sus sequence 3'OH-UGCUUNUNNUNN

UUUGGU,

where NrepresentseitherG, U,or

C,butneverA.

Fromourpresentknowledge of VSV replica-tion,twofunctionsarethoughttobecodedfor,

atleast inpart, inthisregion. The first function is theinitiation of RNA synthesis,including both genome transcription and replication, and the second function is the nucleation site for the initiation of NC assembly, which has recently been localized within the first 14 nucleotides from the 5' end of the leader RNA(la). On the basis of lesscompleteinformation,wesuggested that the NC assembly signal was a five times repeated A residue ineverythirdpositionatthe 5' endof theleader RNA(thecomplementaryU residues in the genome are underlined above). The addition ofthe (+) leader regions of VSV Ch, Co, Pi to ourcatalog has not contradicted this hypothesis, but rather added further

sup-port. The eighth base from the 5' end of the leader strand, previously an invariant C, has now become an N inthe above consensus se-quence, since this basehas changed in both the Piand Coserotypes. In addition, A residuesare

notfound inthe first 18nucleotides from the 3'

end ofany of the genomes sequenced todate. Thus, the absence of U, the complementtothe signal base A in the NC assembly site of the leader strand, has also been confirmed by the sequencesof theCh,Co, and Piserotypes.

It is difficult to decipher the nucleotides

in-volvedin the initiation of RNAsynthesissimply

by comparing sequences. Within the first 18 bases of all thegenomesandantigenomes

stud-ied,twoblocks of nucleotidesare invariant: 3'-UGCUU(bases1to5)and 3'-UUUGGU(bases 13 to 18). Interestingly, the sequence 3'-UUUGGUisrepeatedatposition31to36of the

three In(-) leader regionsbut not inanyofthe (+) leader regions. Recently, Isaac and Keene (6) and Keeneetal.(9) have shownthat theNS

protein, which is thoughtto play a role in the

initiation of RNAsynthesis, selectively perturbs the methylation of certain nucleotides when

presentonthe NC. Thesefootprint experiments have suggested that the NS protein interacts

with bases 16 to 30 at the 3' end of the (-)

genomeand with bases 17to38atthe 3' end of

the(+) antigenome orthecopy-back defective

interfering genome. These NS interaction re-gions abut the 5' end of the first 3'-UUUGGU sequence onbothgenomesandantigenomes and completely cover the second 3'-UUUGGU on the(+) antigenome. Inaddition, all of thestrong

perturbations on the antigenome were at the ends of the interaction region (bases 17 to38), either withintherepeated hexamersequencesor directly adjacent tothem. Since previous work hassuggested that (-) leaderRNAsare

synthe-sizedmorefrequentlypermole ofgenome

tem-plate than (+) leaderRNAs(14),the hexamer

3'-UUUGGU maybe involved in the initiation of

RNAsynthesis,and itsdoublingatthe 3' end of

the (+) antigenome may explainthis enhanced

rate of (-) leader synthesis. The 3' proximal pentanucleotide UGCUU is probably involved

in boththe initiation ofRNA synthesisand the

initiation of NC assembly.

To determine those nucleotides involved in

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.490.112.394.109.239.2]

terminating the (+) leader RNA chains, we

similarly examined these regions of the (-) genomes fornucleotideconservation. From the

data presented in Fig. 2, we have established that the largest(+)leaderRNAs of VSV Ch and

Pi are47base pairslong, and theCo(+)leader RNA is one or two nucleotides longer. The

nucleotidesequenceswhichappear to havebeen

conserved near the termination sites are also

underlined in Fig. 4 and leadtothe consensus sequence 3'-Y3_6NUIGAAA, where Y repre-sentspyrimidines andNis eithera U or aG. The arrowshowsthe presumed point ofchain termi-nation. Keene etal. (8) havepreviouslypointed

outthat(-) leader RNAsappear tobe terminat-edby thesequences 3'-CCCIAGAA.Thus (+)

and (-) leaderRNAtermination sequences are

related but somewhat different, and this

differ-encemayreflect the relativeeasewith which the viral polymerase reads through these

termina-tionsignalsduringgenomereplication.

In conclusion, extensive conservation of the

nucleotidesequencein theleader RNA

(noncod-ing) regions of five

different

VSV serotypes provides suggestive evidence for the

involve-mentofthese sequences incontrolfunctions. In

addition, the lengths ofthe (+) leader RNAs appear to havebeenhighly conserved,

suggest-ing that the distances between the concerned regions may also play arole in the control of VSV transcription and replication.

ACKNOWLEDGMENTS

Wethank Rosette Bandelier andColettePasquierfor expert technicalassistance.

This work wassupportedbyagrantfrom the Swiss National ScienceFoundation.

LITERATURE CITED

1. Banerjee, A. K., G. Abraham,and R.J. Colonno. 1977. Vesicularstomatitis virus; mode oftranscription (a

re-view). J.Gen.Virol. 34:1-8.

la.Blumberg, B. M., C. Giorgi, and D. Kolakofsky. 1983. VSVNprotein selectivelyencapsidatesVSV leader RNA

in vitro.Cell 32:554-567.

2. Blumberg,B.M.,and D.Kolakofsky.1981.Intracellular vesicular stomatitis virus leader RNAs are found in nu-cleocapsidstructures.J.Virol. 40:568-576.

3. Blumberg,B.M.,M.Leppert, and D.Kolakofsky. 1981. Interaction of VSV leaderRNAandnucleocapsid protein may control VSVgenomereplication. Cell23:837-845.

4. Colonno,R. J.,and A. K. Banerjee. 1977.Mappingand initiation studiesontheleaderRNAof vesicular

stomati-tis virus.Virology 77:260-268.

5.Donis-Keller, H., A. M.Maxam, and W.Gilbert. 1977. Mapping adenines, guanines and pyrimidines in RNA. Nucleic Acids Res. 4:2527-2538.

6. Isaac, C. L., andJ.D. Keene. 1982. RNA polymerase-associated interactions neartemplate promoter sequences of defective interfering particles ofvesicular stomatitis virus. J.Virol. 43:241-249.

7. Keene, J. D., M. Schubert, and R. A. Lazzarini. 1979. Terminal sequences of vesicular stomatitis virus RNAare

both complementary and conserved. J. Virol. 32:167-174. 8. Keene, J. D., M. Schubert, and R. A. Lazzarini. 1980. Intervening sequence between the leader region and the nucleocapsid geneofvesicularstomatitisvirus RNA. J. Virol. 33:789-794.

9. Keene, J. D., B.J. Thornton, andS.U. Emerson.1981. Sequence-specificcontactsbetween the RNApolymerase ofvesicular stomatitis virus and the leader RNA gene. Proc.Natl. Acad.Sci. U.S.A. 78:6191-6195.

10. Kolakofsky, D. 1982.Isolation of vesicularstomatitisvirus defective interfering genomes with different amounts of 5'-terminalcomplementarity. J. Virol. 41:566-574. 11. Kolakofsky, D.,and B. M.Blumberg. 1982. A model for

the control of non-segmented negative strand virus genomereplication,p.203-213.InB. W. J.Mahy,A.C. Minson, and G. K.Darby (ed.),Viruspersistence, sym-posium33.Societyfor GeneralMicrobiologyLtd. Cam-bridgeUniversityPress.

12. Krupp, G.,and H.J.Gross.1979.RapidRNA sequenc-ing: nucleases fromStaphylococcusaureusand Neuros-pora crassa discriminate between uridine and cytidine. Nucleic Acids Res. 6:3481-3490.

13. Laemmli, U. K. 1979. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature(London)227:680-685.

14. Leppert, M., andD.Kolakofsky.1980.Effect ofdefective interfering particles on plus- and minus-strand leader RNAs invesicular stomatitis virus-infected cells. J. Virol. 35:704-709.

15. Leppert,M.,L.Rittenhouse, J.Perrault,D. F.Summers, andD. Kolakofsky. 1979. Plus and minus strand leader RNAsinnegativestrandvirus-infectedcells.Cell 18:235-245.

16. ObUeski,J. F.,D.H. L.Bishop,F. A.Murphy,andE. L. Palmer.1976.Structural proteinsof La Crosse virus. J. Virol. 19:985-997.

17. Peattie,D. A.1979.Direct chemical methodfor sequenc-ingRNA. Proc.Natl. Acad.Sci. U.S.A. 76:1760-1764. 18. Repik, P., A. Flamand, H. F. Clark, J.F.Obijeski, P.

Roy,andD. H. L.Bishop. 1974.Detection ofhomologous RNA sequences among six rhabdovirus genomes. J. Virol. 13:250-252.

19. Schubert,M., G.G.Harmison, J.Sprague,C. S.Condra,

and R. A.Lazzarni.1982.In vitrotranscriptionof vesicu-larstomatitisvirus: initiation with GTPat aspecificsite

withintheNcistron.J. Virol.43:166-173.

20. Semler,B.L.,andJ. J.Holland. 1979. Persistent vesicular stomatitis virus infection mediates base substitutions in viral RNAtermini.J.Virol.32:420-428.

21. Wagner,R. R.1975.Reproductionofrhabdoviruses,p.

1-94.InH.Fraenkel-Conrat andR.Wagner(ed.), Compre-hensive virology, vol.4.PlenumPublishing Corp.,New York.

on November 10, 2019 by guest

http://jvi.asm.org/