Mouse hepatitis virus A59: mRNA structure and genetic localization of the sequence divergence from hepatotropic strain MHV-3.

(1)

0022-538X/81/090823-12$02.00/0 Vol. 39,No. 3

Mouse

Hepatitis Virus A59:

mRNA

Structure

and Genetic

Localization

of the

Sequence

Divergence from Hepatotropic

Strain

MHV-3

MICHAEL M. C. LAI,* PETER R. BRAYTON, ROBERT C. ARMEN, CHRIS D. PATTON,

CHARLES PUGH,tANDSTEPHEN A. STOHLMAN

Departmentsof Microbiology and Neurology, University of Southern California School ofMedicine, LosAngeles,California90033

Received 9April 1981/Accepted 26 May 1981

The composition and structure of the mouse hepatitis virus (MHV)-specific

RNA in actinomycin D-treated, infected L-2 cells were studied. Seven

virus-specific RNA species with molecular weights of0.6 x 106,0.9x 106,1.2 x 106,1.5

x 106,3.0 x 106,4.0x 106,and 5.4x 106 (equivalent to the viral genome) were

detected. T1oligonucleotide fingerprintingstudies suggested that the sequences

of each RNA speciesweretotally includedwithin the next larger RNA species.

Theoligonucleotidesof each RNAspecies were mapped on the 60S RNA genome

ofthe virus. Each RNAspecies contained the oligonucleotides startingfrom the

3'end of thegenomeandextending continuously for various lengths in the 3'

--5' direction.Allof the viral RNA species containedapolyadenylate stretch of100

to 130 nucleotides and probably identical sequences immediately next to the

polyadenylate. These data suggested that the virus-specific RNAs are mRNA's

and have astairlike structure similar tothat of infectious bronchitis virus, an

avian coronavirus. Aproposal is presented, basedonthemRNAstructure, for the

designation of thegenes onthe MHVgenome.Using thisproposal, thesequence

differences between A59, a weakly pathogenic strain, and MHV-3, a strongly

hepatotropic strain, werelocalizedprimarilyinmRNA's 1 and3,corresponding

togenesAand C.

Murine hepatitis viruses (MHVs) are

mem-bers of theCoronaviridae,whichinclude viruses

isolated frommanyspecies ofanimals (10, 18).

They contain a positive single-stranded 60S

RNAgenomewithamolecularweight of5.4 x

106 (6, 7, 19).Several other coronaviruses have

been reportedto have a similar genome

struc-ture(4, 8, 9).Aviralgenomeof this sizecancode

for all of the MHV structural proteins, which

includegpl80/90, pp5O,andgp23 (17), andmay

also haveenough informationtocode for several

nonstructural proteins (6). However, no

at-tempts have yet been made tolocalize the

ge-netic regions which code for thepossible gene

products.

Our laboratorieshavebeen interested in

iden-tifyingthegenetic regionsof the MHVgenome

associated with various viralproperties,

partic-ularly those responsible for the ability of the

virusto cause different diseases. Recently, we

have shown thatthegenomicRNAsofaweakly

pathogenicstrain, A59, andastrongly

hepatitis-inducing strain, MHV-3, have verysimilar

oli-gonucleotidefingerprintsexcept forafew

oligo-nucleotides (7). We have also found that a

de-myelination-inducing MHV strain, JHM-DS,

and aJHM strain whichcauses very little

de-myelination, JHM-DL, differ in onlyafew

oli-gonucleotides (S. A. Stohlman, M. M. C. Lai,

and L.Weiner,unpublisheddata). These viruses

thusprovideapotentialsystemfor the

identifi-cation of genetic regions responsible for viral

pathogenicity. As an initial step toward this purpose, we studied the structure of the

intra-cellular viral RNAs in order to identify the

mRNA specieswhicharetranscribedfrom such

geneticregions.

Several earlier reports have

attempted

to

characterize thevirus-specific mRNA in

MHV-infected cells (11, 12). In these

studies,

only

a

few RNA species of 18 to 28S were detected.

These RNAsare notlikelytorepresent the total

RNA species transcribed from the MHV

ge-nome. Recently, Stern and Kennedy (14, 15)

foundsix InRNAspeciesinthe avian

coronavi-rus-infected cells. Furthermore,

they

showed

thatthesemRNA's hadanestedstrucutre.

Sev-eral laboratorieshavealso foundseven mRNA

speciesinMHV-infectedcells(13;J.

Leibowitz,

personal communication). In this report, we 823

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

824 LAI ET AL.

show that themRNA's ofsimilarnature existin MHV-A59-infected cells.We have further

char-acterized their structure.

MATERIALS AND METHODS

Viruses andcells. The A59 strain of MHVwas

originallyobtained fromC.Bond,of theUniversityof California at SanDiego. This strain hasbeenshown tobeweakly pathogenic (11).Thehepatotropic MHV-3strainwasobtainedfrom M.Collins,of Microbiolog-ical Associates, Bethesda, Md. The L-2 cells were

obtained from L. Sturman,New YorkState Depart-mentofHealth,Albany,N.Y. The DBT cell linewas

obtained from K.Fujiwara, UniversityofTokyo,

Ja-pan.Viruseswereplaque purifiedonDBT cellsbefore

use.

Radiolabeling and purification of virus. For preparation of 32P-labeledviralRNA,theviruseswere grown onmonolayers of DBTcells in 15-cm dishes. Viruswasadsorbed atamultiplicityof infection ofone

tofive at370Cfor1h. Afteradsorption,the inoculum

wasremoved,and 25 ml of Dulbeccomodified

mini-mum essential media (DMEM) containing 1% heat-inactivated (56°Cfor 30min)fetal calfserum (FCS) wasadded.After 4 h at37°C,themediumwasremoved andreplacedwithphosphate-free DMEM which

con-tained 1% dialyzedFCS and 200,uCi of

'Pi

perml (ICNPharmaceuticals). Supernatantfluidswere col-lected 8 h later and clearedofcelldebrisby centrifu-gationat15,000x gfor30minat 40C. Viruswasthen pelletedandpurifiedbysedimentationon linear20 to 42%sucrosegradientsaspreviouslydescribed(6).The 32P-labeledRNAwasextractedfrom thepurifiedvirus accordingtopublished procedures (7).

Preparation of intracellular virus-specific RNA. DBT and L-2cells wereused toprepare intra-cellular virus-specific RNA. Both cell lines were in-fected with MHV-A59atamultiplicityofonetofive for 1 h at 370C. After adsorption, DMEM which contained 2% dialyzedFCS and 1 to 3jigof actino-mycinDperml(a giftof MerckSharp&Dohme)was

added.At3to 4 hpostinfection,100,uCiof[3H]uridine

per mlwas added, and thevirus-specific RNAwas

extractedat6 to 7 hpostinfection.L-2cellswereused toprepare32P-labeled RNAbecause theRNA synthe-sis in DBTcells wasmarkedlyreducedwhen thecells were incubated in phosphate-depleted medium. L-2 cellswereincubated for 6to8 h inDMEMcontaining 2%dialyzedFCSand1/10of the normalconcentration ofphosphatebefore infection. Afterinfection,the L-2 cellswerecultured in phosphate-free DMEM which

contained 2%dialyzedFCS,3ugofactinomycinDper

ml,and 250 MCiof'Piperml.Virus-specificRNAwas

extracted from L-2cellsat7to9 hpostinfection. For extraction ofintracellular viral RNA,

mono-layersof infectedcellswerechilledonice and washed

three times withSSbuffer(10mMTris,60mMNaCl, and1mMEDTA, pH8.5).Thecellswerelysed in the

SSbuffercontaining0.5% TritonN-101, and the nuclei

wereremovedbycentrifugationat1,800xgfor5min.

The supernatant fluidwasthen adjustedto contain 1%sodiumdodecyl sulfate (SDS)andwasextracted

at 40Cwith chloroform-phenol (1:1). The RNA was

precipitated bythe addition of 2volumes ofethanolat -20°C overnight.

Initially, the RNA wasanalyzed byvelocity sedi-mentation in linear 15 to 30% sucrose gradients pre-pared in SS buffer (pH 8.0) which contained 0.5% SDS. Centrifugationwasperformed inanSW40 rotor at20,000 rpmfor 16 to 20 h at 200C. Fractions (0.2 ml) were collected withan ISCO fraction collector and were monitored by UV absorbance at 254 nm. 3p_ labeled DBT cellular RNAwasaddedtoprovide 28S, 18S, and4Ssedimentation markers.

Agarose gelelectrophoresis. Two different con-ditions were used for agarose gelelectrophoresis. In mostof theexperiments, the RNAsamples were dis-solved in RE buffer (50 mM boric acid,5mMsodium borate,1mMEDTA,and10mM sodiumsulfate, pH

8.1), heat denaturedby boilingat1000C for1min, and electrophoresed in 1% agarose gel slabs (15 by 20 by 0.3 cm) made in RE buffer at 90 V for 5 h. For determination of molecular weights of RNAs, the RNA samples were dissolved in RE buffer and electro-phoresed in agarose gel slabs made in RE buffer con-taining 5 mM methylmercuric hydroxide (2). After

electrophoresis, the gels were wrapped with cello-phane andexposed to Kodak XR film with an

inten-sifyingscreenfor variousperiods of time.

For preparativegelelectrophoresis, the RNA sam-pleswereelectrophoresedin agarosegels with or with-out methylmercuric hydroxide as described above. The RNA bandswereexcised from thegel, suspended in 2 to 3 ml of distilled water, and crushed with a Dounce homogenizer. For RNAs excised from the

methylmercury gels,the gel sliceswerefirst washed withtwochanges of 0.5 Manmmoniumacetatebefore homogenization.After centrifugation at 20,000xgfor 15min to remove thegel, the supernatant was adjusted to 0.15M ammonium acetate and precipitated with 2.5 volumes of ethanol. This extraction procedure removed at least 80% of the radioactivity. Further elution of thegel didnotremoveadditional RNA. The RNAwasprecipitated by sedimentation at 20,000xg for15min andredissolved in0.4ml of 0.2 M ammo-niumacetate. The RNA samples were again

centri-fugedat20,000xgfor15mintoremovecontaminating

gel pieces. The RNA in thesupernatantwasfinally precipitated with2.5volumes ofethanol.

Oligonucleotide fingerprinting.The32P-labeled

RNAextracted from purified virions or eluted from

gelswasdigestedwithRNaseT,and analyzed by two-dimensionalpolyacrylamide gel electrophoresis

essen-tially as previously described (7). Briefly, the first dimension was performed on 10% polyacrylamide-0.125% bisacrylamide gel slabsin citrate buffer, pH 3.3,containing6M urea.Electrophoresis was carried out at 700 V for 4 h. The second dimension was performedon22%polyacrylamide gel-0.15% bisacryl-amidegel slabs in Tris-borate buffer, pH 8.0, at 600 V for16h. Afterelectrophoresis, the gels were wrapped withcellophane and exposed to Kodak BB-1 film with anintensifyingscreenat40C.

Oligonucleotide mapping. The 32P-labeled 60S RNAwasdissolved in 0.3 ml of 0.05 M sodium carbon-ateandincubated at 50°C for 1 to 5 min. The reaction mixture was then neutralized with 15

pl

of 1M acetic acid andprecipitated with 2.5 volumes of ethanol. The RNAwaspelletedby sedimentation at 20,000xg for 15minandwasthen fractionated by oligodeoxythym-idylate[oligo(dT)]-cellulose column chromatography J. VIROL.

http://jvi.asm.org/

(3)

MHV-A59 mRNA STRUCTURE 825

accordingtotheprocedure ofAvivand Leder (1). In thelater partof theexperiments, the alkali digestion oftheRNAwasomitted. Instead,the total RNA from purified virus, without prior separation by sucrose

gradient sedimentation, was used directly for oligo(dT)-cellulose chromatography. The RNA

pre-pared by this procedure was degraded enough for oligonucleotide mapping. The polyadenylate [poly-(A)]-containing RNAselected by oligo(dT)-cellulose chromatography wassedimented ina 10 to25% su-crosegradient made in0.01MTris-hydrochloride,pH 7.4, 0.01 MNaCl, and 0.05% SDSinanSW41rotor at

40,000rpmfor 4.5 h. TheRNAs of differentsizeswere

pooled separatelyandwereused forT, oligonucleotide fingerprinting.

Basesequenceanalysis.Thepoly(A)-containing oligonucleotides separated by two-dimensional poly-acrylamide gel electrophoresiswereexcised from the

gels,eluted with 0.5 ml of 0.5 MNaCl, and precipitated with2.5volumes ofethanol. The oligonucleotideswere

pelleted by sedimentationat 20,000 xgfor 15mi, redissolvedin10plof waterwhichcontsinedRNase A at0.1 mg/ml, and incubated at370Cfor 30 min. Afterincubation, thedigestswerespottedon DEAE-cellulosepaperandelectrophoresedin acetic acid-pyr-idine buffer(pH 3.5)at1,500Vuntil the xylene-cyanol dyehad traveled 15cm. TheDEAE-cellulose paper wasdried andexposedtoKodakBB-1 film with inten-sifyingscreens. The basecompositionofthe oligonu-cleotideswasdeterminedby comparisonwith nucleo-tide standards obtainedbydigestionof the total

un-fractionated RNA(see Fig. 7) (3).Eachspotwascut outof the paperandwas counted in toluene-based scintillation fluid. The relative ratio of eachnucleotide

wasdeterminedfrom theradioactivitypresentineach spot.

RESULTS

Kinetics ofintracellular viral RNA

syn-thesis.MHV infection of L-2orDBT cells did notresult in inhibitionof hostcell macromolec-ular synthesis (datanot shown). Therefore, to

study the intracellular viral RNA, the RNA

synthesisof the host cellswasinhibitedby

ad-dition ofactinomycin D. We found that, with

actinomycinD ataconcentrationof 1

,ug/ml

for

DBT cells and 3 pg/mlfor L-2 cells, the host

RNA synthesis was completely inhibited. At

these concentrations, no appreciable effect of

actinomycinDonviralreplicationwasnoticed, asjudged bythe kinetics of virusgrowth,virus

yield, and cytopathic effects. Similar

observa-tionshave also beenmadebyRobb and Bond

(11). However, athigherconcentrations of

acti-nomycin D (5 pg/ml), the yield of infectious

virus from DBT cellswasreducedby 100-fold.

L-2cellswere moreresistanttoactinomycin D, andtherewas noappreciablereduction in virus

yield even at 5

,ug/ml.

Wetherefore used

acti-nomycin D at 1 ug/ml for DBTcells and at 3

Ag/ml

for L-2cellsthroughoutthisstudy. Wefirst determined the kineticsof the

intra-cellular virus-specific RNA synthesis in DBT

cells. Thecells wereinfected with MHV-A59at

a multiplicity of one to five and were labeled

with [3H]uridine for 1 h atvarious times after infection. Table 1 shows therateof viral RNA synthesis and virus yieldsatdifferent time points postinfection. Afteralag of2h, thevirus-specific

RNA synthesis reached a peak at 6 to 7 h

postinfection. Thereafter, therateof viral RNA synthesis declined. Bycontrast, the titer of the infectious virus released into the medium did

notreach its peak until 8to 9 hpostinfection. For most ofthe studies, we used intracellular

viral RNA whichwasprepared during maxinum

intracellular RNA synthesis. Studies performed with L-2 cells cultured in thepresenceof 3 pigof

actinomycin Dpermlyielded similar data

(Ta-ble1), with the only exception being that

maxi-mumRNAsynthesiswasdelayed by1 h inL-2 cellscompared with DBTcells.

Sucrosegradient sedimentation and elec-trophoretic analysis of MHV-specific RNA. To determine the species and structure ofthe virus-specific RNA in the MHV-infected cells, the A59-infectedDBTorL-2cellswerelabeled

with[3H]uridineor

'Pi,

and the total

cytoplas-TABLE 1. Rate of MHV-A59virus-specificRNA

synthesisin DBT and L-2 cells

Time post- Radioac- Virus yield(PFU/ml)' Cella infection tiVityb

(h) (cpm) -AMD +AMD

DBT 2 0 1x 102 8x 101

3 650 4x10( 2 x103

4 2,700 2x104 3x104

5 4,600 5 x104 5 x 104 6 23,000 4 x105 4 x105 7 11,200 3x106 3 x105

8 2,850 4 x106 1x106

9 1,470 5x106 2x106

L-2 4 0 2 x103 2x104

6 680 3x103 2 x104 7 1,320 1x104 2x105

8 3,620 2 x105 4 x106

9 510 3 x106 8x106

10 480 2 x106 3 x105

11 420 8x105 4x105

aDBT

cells

were treated with actinomycin D (AMD)at 1ug/mland[3H]uridineat100pCi/ml.L-2 cellsweretreated withactinomycinDat3pg/mland

[3H]uridineat 10pCi/ml.Thecellswerelabeled for1

h atthe specified timepostinfection.After labeling,

thecellsweredisruptedwith 1%SDS,andtheamount

of[3H]uridineincorporation wasdeterminedby

pre-cipitationwithtrichloroacetic acid.

b The amount of trichloroacetic acid-insoluble

[3H]uridine

incorporationafterlabelingfor1h inthe A59-infected cells in the presence ofactinomycinD. Counts of the uninfected cellshavebeensubtracted.

'DeterminedbyplaqueassayonDBT

cells.

http://jvi.asm.org/

[image:3.499.255.448.366.537.2]

(4)

826 LAI ETAL.

mic RNAwassubjectedtosucrose

gradient

sed-imentation. At least fourRNAspeciescould be

detected: 60S, 32S, 22S, and 18S (Fig. 1). The

60S RNA cosedimented with the

genomic

RNA

extracted from thepurifiedvirus(Fig. lb).

Fur-thermore, the T1 oligonucleotide fingerprint of

this RNA wasidentical to that of the 60S viral

genomicRNA(seebelow).Therefore,thisRNA

species represented the genomic RNA of the

virus. The rest of the RNAs

represented

the

virus-specific subgenomic

RNAs,

since the

mock-infected,

actinomycin

D-treated cellsdid

notcontain anyRNA of this size

(Fig.

la). An

additionalspeciesofRNA, 4S,

likely

represented

28S 18S 4S

a)

10 20 30 40 50 60 FRACTIONNUMBER

70

thetRNA's whichwerefoundin theuninfected

cellsaswell. ThisRNA

species

wasnot

further

characterized.

The

virus-specific

RNAs were further sepa-rated

by

electrophoresis

on1%agarosegels.Six

virus-specific

RNA bandscould beresolved

(Fig.

2a).

Longer

exposureof the

gels

didnotreveal

anyadditional RNA bands. When these RNAs

were

analyzed

by

electrophoresison

denaturing

gels

whichcontainedmethylmercuric

hydroxide

(2),

essentially

similarresultswereobtained. Six RNA

species

were

consistently

resolved

(Fig.

2b).

However,

in a few RNA preparations, an

additional RNA species was found (Fig.

2c).

These

findings

confirmed the datarecently

ob-tained in other laboratories with the

A59-in-fected Sac (-) or 17 CI-1 cells,whichalso

con-tained sevenvirus-specificRNAspecies (13; H.

Wege,

S.

Siddell,

M.Sturm,andV.ter

Meulen,

in V. ter

Meulen,

S.

Siddell,

and H.Wege,

ed.,

Biochemistry

and

Biology of Coronaviruses,

in

press; J.

Leibowitz,

personal

communication).

We

designated

theseRNAs,inorder of

decreas-ing size,

RNAspecies1through7.Themolecular

weights

ofthese RNAspeciesweredetermined

from the

migration

rates of the RNAs in the

agarose

gels containing

methylmercuric

hydrox-ide

(Fig. 2b,

c). Under these conditions, the

migration

rate of the RNA was inversely

pro-portional

to the logarithm of the molecular

18S

-C)x

0.

CV FRACTION NUMBER

FIG. 1. Sucrose gradient sedimentationof the vi-rus-specific RNA intheMHV-infectedcells. [3H]ur-idine-labeled RNA was extracted from the

MHV-A59-infectedandactinomycin D-treatedDBTcellsat

7hpostinfection andthen sedimentedthrougha 15

to30%linearsucrosegradient containing0.5%SDS inanSW40rotorat20,000rpmfor20h. (a) The

32p-labeled RNA from the mock-infected and

acti-nomycinD-treated DBTcellswascosedimented. (b)

The32P-labeled 60S RNAextractedfromthepurified MHV-A59 virionwasincluded forcomparison. The

positionsofthe28S, 18S, and4SmarkerRNAs are

indicated.

FIG. 2. Agarosegel electrophoresis ofthe

virus-specific RNA in the A59-infectedcells. The 32P-la-beled RNAfromtheA59-infected L-2 cellswasheat

denaturedat100°Cfor1min and electrophoresedon

1%agarosegelwithout(a)orwith (band c) methyl-mercurichydroxideat90Vfor5h.TheRNAs in(a), (b), and (c) were from differentpreparations. The

positionsofthe 35SRoussarcoma virusRNAand 28S and 18SrRNA's areindicated. The gels in(b)

and(c) wereoverexposedto reveal the lesserRNA species.

12

10_

I

1 8 I

X 6

-0~

4

2

_ v

I

o

0

x M 4

C) I

J. VIROL.

http://jvi.asm.org/

[image:4.499.57.252.225.528.2] [image:4.499.279.426.389.546.2]

(5)

weight (2). From these data, the molecular

weightsoftheintracellular virus-specific RNAs were determined to be5.4 x 106, 4.0 x 106, 3.0

X 106¶ 1.5X 106, 1.2 x 106,0.9 x 106,and 0.6 x

106 for RNAspecies1through7,respectively.

Itisnoteworthy that the relative ratio of these

RNAspeciesremainedquiteconstant,withthe

exception of RNAs1and4.RNA1 (equivalent

totheviral genome)wasthemostvariable

spe-cies. Insomepreparations,itwasthe

predomi-nantRNAspecies (Fig. 2b, c).Itisprobable that

the ratio of all the subgenomic RNA species

remainedconstant,whereas the increase ofRNA

1 was due to accumulation of genomic RNA

destinedtobeincorporated into the virion. RNA

4 waspresentinthe lowestamountin allof the RNA preparations. In L-2 cells,this RNA spe-cieswasusuallynotdetectable.

Oligonucleotide fingerprintingof the

vi-rus-specific RNAs. Thesumofthe molecular

weights of the six subgenomic RNAs exceeded

the genome size, suggesting that these RNAs

overlap in their sequences. To determine the

structural relationship between the various

RNAspecies,oligonucleotide fingerprintsof

ev-ery virus-specific RNA, except RNA 4, which

was usuallynotdetectable inoursystem, were

examined. The 32P-labeled RNAs were eluted fromagarosegels, digested with RNase T1, and

separated by two-dimensional polyacrylamide gel electrophoresis.The T1 oligonucleotide

fin-gerprints became increasingly complex as the

size of the RNAs increased (Fig. 3).

Further-more,mostof theoligonucleotidespresentin the

small RNAspecieswereincluded in the

finger-prints of thenextlarger RNAspecies.Also,most

of the oligonucleotides in each RNA species

were also presentin the 60Sgenomic RNAof

MHV-A59, suggestingthattheyrepresented

se-quences of the same polarity as the genomic

RNA. The RNA 1 had an oligonucleotide

fin-gerprint identical tothatof the genomic RNA, confirmingthat RNA 1representsthe

intracel-lular form of the genomic RNA. Finally, every

RNA speciescontaineda

poly(A)

spot

(located

at the lower left comers of the

fingerprints),

indicating that they are

polyadenylated

and

probablyfunctionasmRNA's.

The oligonucleotide fingerprintsofseveral of

the RNAspeciesalso containedsomeanomalous

oligonucleotides: (i) RNA 7containedan

oligo-nucleotide, 19, which was present only very

weaklyor not at allinothersubgenomic RNA

species;(ii)RNA 6containedan

oligonucleotide,

19a, and RNA 5 contained anoligonucleotide,

3a, both ofwhichwere not present in theviral

genome,nor weretheydetectedinother

subge-nomic RNA species; (iii) RNAs 2 and 3 were

slightly contaminated with the degradation

products ofRNA 1; however, thecontaminating

oligonucleotides could bediscerned due to the muchlowermolar ratioatwhich they occurred (Fig. 3). It appeared that the amounts of some

of the oligonucleotides, e.g., oligonucleotides 3

and5in RNA 3, were reduced in RNA 2. The

origin of these anomalous oligonucleotides is

currently beingstudied.Itispossible that these

oligonucleotidesrepresent the extreme 5' ends

of the subgenomicRNA species.

Mapping of thesubgenomic RNA species

onthe RNAgenome. Theoligonucleotide

fin-gerprints of the virus-specificRNA species

sug-gested that the MHV subgenomic RNAs had

overlapping sequences, with the sequences of

the smaller RNA being contained within the

next-larger RNA species.However, these studies

didnotreveal the genetic localization of these

mRNA'sonthe RNAgenome.Itwasimportant

to determine the genetic regions from which

each mRNAspecieswastranscribed. To

accom-plish this,we studied themapposition of each

oligonucleotideonthe RNA genome. The

par-tially degraded 3P-labeled 60SRNAofA59 was

fractionated by chromatography on

oligo(dT)-cellulose. Thepoly(A)-containing RNAwas

se-lected and fractionated bysucrose gradient

sed-imentation. The genomic RNA from A59 was

degradedtosuchanextentthat thepeak size of

the RNAwas188 (Fig. 4). The

poly(A)-contain-ing RNAs of different sizeswerepooledasshown

in Fig. 4 and analyzed by T1 oligonucleotide

fingerprinting. Several of the RNAfingerprints

are presented in Fig. 5. All these RNAs

con-tained the sequences starting from the 3' end

andextending for various distance toward the 5'

endof thegenome.Therefore,therelative

posi-tion ofeacholigonucleotideonthe RNAgenome

could be arranged in several groups from 3'

toward the5' end.Furthermore,thepositionof

theoligonucleotidesonthe RNAgenomecould

be estimated from the smallest size of the

poly(A)-containing RNA in which the

oligonu-cleotidewasdetected. Bythisapproach,all the

oligonucleotides of thegenomicRNA of MHV-A59 were mappedin severalgroupsonthe

ge-nome (Fig. 6). By comparison with this

oligo-nucleotide map, it.was found that all of the

oligonucleotidesinRNA7 werelocatedatthe3'

end of thegenome. Therestof thevirus-specific

RNAs alsocontained theoligonucleotides start-ing from the 3' end and extending for various

lengths toward the 5' end (Fig. 6). This result

suggested thatthesubgenomic RNAsof

MHV-A59hada structuresimilartothatproposedfor

avian infectiousbronchitisvirus (14, 15).

How-ever, itwas also clear from Fig. 6 that several

RNAspecies,including2, 3, 5,and6, contained

the oligonucleotides which either were absent

http://jvi.asm.org/

(6)

828 LAI ET AL.

A.

:E0-

A0

~

0

B o0

C

00

00500 0

OO 10 X52

37. C-O44~~ 1

350 4342 1 ~ 3

340 16 822

00 10

[image:6.499.60.450.74.572.2]

14 13 <1f05

FIG. 3. Oligonucleotide fingerprintingof the intracellular A59-specific RNAs. The virus-specific RNAs wereelutedfromagarose gels, digested with RNaseT1, and separated bytwo-dimensionalpolyacrylamide gel electrophoresis. Electrophoresiswasfromleft to right in the firstdimension and from bottom to top inthe second dimension. The numbering of the spots is as indicated in the schematic sketch (B). Only the oligonucleotides whichappearedforthefirst time in the mRNA's of increasing size were numbered. The numbersattheupperleftcornersofeach panel represent the RNA species referred to inFig.2.Abbreviations: V,viriongenomic RNA;A, poly(A).

from theviral genome or were not contiguous nome.Whether theseoligonucleotidesrepresent

with therestof theoligonucleotides onthege- splicing of the sequences is unknown at the

J. VIROL.

http://jvi.asm.org/

(7)

I A1 - 1 -

HF--25 -28S

CM 35S

20

'-l 60S

x

E 15

0

X 10

5

0 5 10 15 20

Fraction Number

FIG. 4. Sucrosegradientsedinentationof the

par-tially degraded poly(A)-containing A59 genomic RNA. Thepoly(A)-containingfraction fromthe

par-tially degradedA59RNAwasselectedby

oligo(dT)-cellulose column chromatographyand sedimentedon

a 10 to25%sucrosegradient made in 0.01MTris hydrochloride,pH7.4, 0.01MNaCI,and 0.05% SDS

inanSW41 rotorat40,000rpmfor4.Eh. TheRNA

ofdifferent sizeswaspooled into the fivefractionsas

indicated.

presenttime.

Sequences immediately adjacenttothe 3' ends of thevirus-specific RNAs.The oligo-nucleotidemappingstudiessuggestedthatallof

the virus-specific RNAs contained the oligo-nucleotides starting from the 3' end of the

ge-nome. Thus, they should have identical 3' end

sequences.However,this conclusionwas

uncer-tain since severalRNAspeciescontained oligo-nucleotidesnotpresentin thegenome,andsome

of theoligonucleotidesinthe smaller RNA spe-cieswerenotpresentin thelargerRNAs. Itwas

therefore possible that some of these RNAs

might have different 3'-endsequences. To test

this possibility, the poly(A) spotsfrom the

oli-gonucleotide fingerprintsof each RNA (Fig. 3)

wereisolated, digestedwith RNaseA,and sep-aratedbyelectrophoresisonDEAE-cellulose

pa-peratpH 3.5. Thesepoly(A) spotsshould

con-tain not only the poly(A) tracks but also the

sequences which are immediately adjacent to

the poly(A) and extend inward until the first guanineresidue.Figure7shows suchananalysis

for RNAs 5, 6, and 7. From the molar

yields of each base, the base composition of

the sequences immediately adjacent to the poly(A) spots was determined to contain C(AC)(AG)(A2U)(A3C)(A4N)Alol-13o (Table 2).

It canbe seen that all of the RNAs contained

identical base compositionsandlikely had

iden-tical sequences at the 3' end. These data also

show that all of the RNA species contained a

poly(A) stretch of 100 to 130 nucleotides long.

The presence of (AG) residues within this

stretchcannotbe explainedat thepresenttime.

It was reproducible and may represent

se-quencesprotected from RNase T1 digestionby

the poly(A) stretch. Similar studies have been

performed with thepoly(A) spots of the RNAs

1, 2, and 3. All ofthem appeared to have the samebase sequences,although the molar yield

of each base was not confirmed by counting

radioactivity because ofthe low counts

associ-ated with each spot (data not shown). These

data thereforesuggestthatallthevirus-specific

RNA species contained identical 3' end

se-quences.

Genetic localization of the sequence di-vergencebetween A59and MHV-3.Wehave

previouslyshown that A59,aweakly pathogenic

strain, and MHV-3, a strongly hepatotropic

strain, haveverysimilaroligonucleotide

finger-printsexceptfortwounique oligonucleotidesin

MHV-3 and four in A59 (7). Some of these

oligonucleotides might be derived from the

ge-netic regionsresponsible for the biological

dif-ferences between these two virus strains. We

therefore attemptedtolocalizethese

oligonucle-otides on the RNA genome. As shown

previ-ously,A59contains fouruniqueoligonucleotides,

9, 13,19,and 28, whicharenotpresentin

MHV-3 (7). Comparison of Fig. 3 with Fig. 5 shows thattwoofthe A59-specific oligonucleotides, 9 and13, were presentinthegeneticregionat6 to

10kilobases from the 3' end which corresponds

toRNA3(Fig.6). Incontrast,thegenetic

local-ization of theoligonucleotides 19 and 28 is less

certain. Although oligonucleotide 19 appeared

tobepresentin RNA7, itwasabsent in RNAs 5and6. Oligonucleotide28appearedalsotobe presentin RNA3 (Fig. 3),but itwasfoundby fingerprinting degradedvirion RNAtobe closer

tothe 5' end ofthegenome (Fig. 5).This

oligo-nucleotide was present at lower molar yields

than other

oligonucleotides,

thusincreasingthe

difficultyofprecise

genetic

mapping.

The

oligon-ucleotide 28 was therefore not

represented

in

Fig.6.

We alsomappedthetwo

MHV-3-specific

oli-gonucleotides,aandb,onthe RNAgenomeby

fingerprinting the

poly(A)-containing

RNA of

different size fractions.

Figure

8showssomeof

the representative

fingerprints.

By

this

ap-proach, itwasdeterminedthat a waslocalized

at 6 to 7 kilobasesfrom the 3'

end,

whereas b

waslocalizedvery closetothe 5' end.Thesetwo

regions correspond to the

regions

contained

within RNA 3 and RNA 1,

respectively.

From

http://jvi.asm.org/

[image:7.499.53.244.62.264.2]

(8)

[image:8.499.60.445.74.562.2]

J. VIROL. 830 LAI ET AL.

FIG. 5. T1-oligonucleotidefingerprints of the poly(A)-containingA59 RNAsofdifferent size fractions. The RNAscollectedasindicated inFig.4wereanalyzedbyT,oligonucleotide fingerprinting.Oligonucleotides whichappeared for the first time intheRNAsof increasing sizewerenumbered. The numbering of the total largeT1oligonucleotidesisshownforthe60S RNAand isidenticaltothat inFig.3(B).

the genetic mapsof A59 and MHV-3, we

con-cluded thatatleast RNAs3and1contained the

geneticdifferencesbetween thesetwostrains.

DISCUSSION

Our resultsshowthat L-2 cellsinfectedwith

the A59 strain of MHV contain six to seven

http://jvi.asm.org/

(9)

MHV-A59 mRNA STRUCTURE 831 608

4

45S

4

35S 28S 18S 15S

4

18 16 14 12 10 8 6 4 2 0

5D

I i I i - tl%3

5'~

~~i

1[1[@

§ § | 1

[12 ][4,30,

][,40 31

33,31,21,7 4 22337 19

mRNA

(19a) # 7

(16) (5.15,11) .#, *2

w*#2

VA

#1

FIG. 6. Oligonucleotidemap of the A59 genome and its corresponding subgenomic RNAs. TheT1 oligo-nucleotides of the A59genomic RNA (numbered as shown in Fig. 3 and 5) were arranged into several groups ondifferent regions ofthe genome. Since the genomic RNA has a molecular weight of 5.4 x1iO,thegenome length is assumedto be 18kitobases. The subgenomic RNAs are also presented. The oligonucleotides in parenthesesondifferent subgenomicRNAs either wereabsentfromthegenomic RNA or were not localized in the corresponding regionsof the genome. Oligonucleotides in common with those on the genomic RNA were not labeledon thesubgenomic RNAs. The circled spots, 9, 13, and 19, arespecificfor A59 andnot presentin MHV-3(7).

FIG. 7. Base sequenceanalysisofthe

poly(A)-con-taining oligonucleotides ofthe A59 mRNA's. The

poly(A)-containing oligonucleotides were isolated from the oligonucleotidefingerprintsofeach virus-specificRNAspecies asshown inFig.3. These oli-gonucleotidesweredigestedwith RNase A and elec-trophoresedonDEAE-cellulosepaperinacetic acid-pyridinebuffer,pH 3.5, at 1,500Vfor4h. The60S RNAdigestedwith RNase A andT,wasincludedas

TABLE 2. Partial base sequenceanalysisofthe 3'-end sequencesa

RNAspecies Basecomposition 5 C(AC)(AG)(A2U)(A3C)(A4N)Aioi

6 C(AC)(AG)(A2U)(A3C)(A4N)Aj13 7 C(AC)(AG)(A2U)(A3C)(A4N)A13o

a The poly(A)-containing

oligonucleotides

were eluted from thetwo-dimensionalfingerprints, digested with RNase A, andseparated by electrophoresis on

DEAE-cellulosepaper atpH 3.5 (Fig. 1).Each spot wascutout, and therelative ratio of each spotwas determinedbycountingintoluene-based scintillation fluid.

virus-specific RNA species. Similar findings

haverecently been obtained with other cell lines,

including Sac (-) and 17CI-1 cells(13;Wegeet

al., in press;J. L. Liebowitz, K. C.Wilhehnsen,

and C. W. Bond, personal communication).

However, in L-2 cells, one of the al., in press;

Liebowitzetal.,personal communication).

How-ever, in L-2cells, oneof the RNAs, 4, wasnot

always detectable or present only in very low

quantity. It is not clear whether this variance represents a host cell-determined differential expression of viral RNA. We have also shown

that all of the virus-specific RNAs contain

poly(A)segmentsof100 to130nucleotideslong.

This is similar to the length of the poly(A)

present in the viral genome, which has been

estimated tobe about 90nucleotides

long

(21).

mono-andoligonucleotidemarkers(3). The relative ratio ofeach spotwasdetermined. Onlythosefrom

the mRNA's 7, 6, and5 areshown here. Abbrevia-tions:A,AMP; C,CMP; G,GMP; U, UMP;C."cyclic

CMP;N,anynucleotidesotherthanAMP.

http://jvi.asm.org/

[image:9.499.52.445.59.200.2] [image:9.499.99.197.289.576.2]

(10)

832 LAI ET AL.

FIG. 8. Oligonucleotidefingerprints ofthepartiallydegraded, poly(A)-containingMHV-3genomicRNA species. The32P-labeled,60SMHV-3 RNAwasbriefly digested,selectedby oligo(dT)-cellulose

chromatogra-phy, and sedimentedbysucrosegradientsedimentationasdescribedforMHV-A59(Fig. 4).Therepresentative

RNAfractionswerepooled andanalyzedbyT,oligonucleotidefingerprinting. Only representativefingerprints

areshown here.(A) 18S;(B) 35S; (C) 50S; (C) 60S. Oligonucleotidesaandbarespecificfor MHV-3 andnot presentin A59(7).

The presence of poly(A) sequences in these RNAs suggests that they represent mRNA's.

This conclusion is consistent with the recent

findings that these virus-specific RNAs were

present in the polysome fractions of MHV-in-fected cells(M. M.C.Lai, C.D.Patton, andS.

A. Stohlman, unpublished data; 13) and that

these RNAscouldserve asmnRNA'sfor in vitro translation (12; S.Siddell,H.Wege,A.Barthel, and V.terMeulen, inV. terMeulen, S.Siddell,

andH. Wege, ed., BiochemistryandBiologyof

Coronaviruses, in press).

Thestructureof these viral mRNA's contains

someveryunusual features.Theoligonucleotide

fingerprinting studies suggest thattheyhave a

"nestedstructure,"astermedbyStern and

Ken-nedy (14, 15); i.e., the sequences of the small

mRNA are contained within the next-larger

mRNA species. Furthermore, the mapping of

theoligonucleotidesonthe genome suggests that

the sequencesof all of thesemRNA'sstartfrom the 3' endof the genome. These data strongly suggest that the structures of the mRNA's of MHV-A59 arevery similartothatof avian in-fectious bronchitisvirus (14, 15). In further sup-port of this proposal, we found that all ofthe

mRNA species contained identical sequences

immediately adjacent to the 3'-poly(A). The

presence ofseveral anomalousoligonucleotides

insomemRNAspecies, however,suggeststhat

MHVmRNA'smightbemorecomplexthan the

infectious bronchitis virus model. Some

oligo-J. VIROL.

http://jvi.asm.org/

[image:10.499.60.456.69.428.2]

(11)

18

5'

I

AI l

16 14

i

70

l

B I

12 10

i F

Il

I

c

8

D

'E

Fi

G

6 4 2 0 Kb

§ I F 3'

| | mRNA

#5

4'

#4

I

#

3

AvAI #2

[image:11.499.45.444.51.200.2]

II

I~~

# 1

FIG. 9. Proposedmedel of the structure of the MHVmRNA's and the corresponding genetic regions on the genome. The proposed seven genetic regions on the MHV genome are based on the data reported here and from other laboratories (13; Wege et al., in press; Leibowitz, personal communication). The boundaries of each gene were estimated by converting the molecular weights of the RNAs into kilobases (kb). Each gene corresponds to the portion ofeach mRNA species which does not overlap with the next-smaller mRNA species.

nucleotideswerepresentin thesmaller mRNA's

butnotinthelargerRNAs,whereasothers were

notpresentin the genome. These findings

sug-gestthatthe 5' ends of these mRNA'scould be

quite unusual. Preliminary data suggestthatall

of them contain thecapsequencesatthe 5' ends.

Weare currently investigating the

oligonucleo-tides whichmightcontain sucha capstructure.

ThemRNA's with suchastairlike relationship

areverysimilar tothose found in retroviruses

and togaviruses (5, 16, 21). In these mRNA's,

usuallyonlytheportionof the sequences which

is unique to a particular mRNA species and

which is not shared with the next-smaller

mRNA istranslated. Ifasimilar relationship of

mRNA'stogenomeoccursduringreplication of

MHV, the MHV RNAgenomecould be divided

intosevengenetic regionscorrespondingtothese

mRNA species (Fig. 9). This model offers a

rational basis for studying the geneticstructure

of the RNA genome of MHV. Thecoding

poten-tials of the mRNA's in another strain of MHV,

JHM, have been studiedbyin vitrotranslation

(12;Siddelletal.,inpress).If these datacanbe

appliedtothe A59strain, itcould suggest that thegeneproductsof thefollowinggenesare:for B,anonstructural

protein

p30;for

C,

gpl80/90;

for DorE, nonstructuralp14; forF, gp23; and

forG, pp6O. In analogytoother virus systems, gene A might code foran RNApolymerase. It remains to beinvestigated whether MHV-A59 codes forsimilarproducts.

The studyof theoligonucleotide maps ofall

of the mRNA species has also allowed us to

localize the genetic differences between two

MHV strains,MHV-3 and A59. Previously, we

have shown that these two strains differ only

slightlyintheiroligonucleotide fingerprints (7).

Since these two strainsare genetically related

butdiffer in their pathogenic

potentials-MHV-3 causes severe hepatitis, whereas A59is only

weakly pathogenic-these oligonucleotide

dif-ferences could suggest the genetic regions

re-sponsiblefor suchpathogenic potentials (7). Our

mapping studies indicate that one of these

re-gions might correspond to mRNA 3 and thus

would reside ingene C, which codes for gp9O/

180,whereas the othercorresponds tomRNA 1

and therefore would reside in gene A, which

might code for an RNA polymerase. Whether

thesetworegionsaredirectlyresponsiblefor the

viralpathogenicityremainstobeinvestigated.If

gene C is indeedresponsiblefor the viral

path-ogenicity, itwill suggestan interesting

mecha-nism for viralpathogenesis:thegeneproductof

thisgene, gp90/180, is themajorenvelope

gly-coprotein in MHV (17) and could determine the host range of the virus. The nonpathogenic

strain, A59, might have mutations inthis gene

and thus becameunabletoinfect thetargetcells.

Further studiesarerequiredtoresolve thisissue.

ACKNOWLEDGMENTS

Wethank T. Hanson for excellent technical assistance and J.Lopez for editorialhelp.

This work was supported in part by grant PCM-4567, awardedbytheNationalScienceFoundation,andbyPublic HealthService grants NS 15079and CA16113,awardedby theNational Institutes of Health.

ADDENDUM INPROOF

We haverecentlyshownthat all of theA59-specific RNAs containidentical5' endsequences, suggesting

synthesisof these RNAs by splicingor another

un-usualmechanism (M.M.C.Lai,C. D.Patton,andS. A.Stohlnan,submitted forpublication).

LITERATURE CITED

1. Aviv,H.,andP.Leder.1972.Purification ofbiologically

active globin messenger RNA by chromatographyin

oligothymidylic acid-cellulose. Proc. Natl. Acad. Sci.

http://jvi.asm.org/

(12)

834 LAI ET AL. U.S.A. 59:1408-1412.

2. Bailey, J. M.,and N. Davidson. 1976.Methylmercury as areversible agent for agarose gel electrophoresis. Anal. Biochem. 70:75-85.

3.Barrell,B.G.1971.Fractionation and sequenceanalysis ofradioactive nucleotides, p. 751-779. In G. L. Cantoni and D. R. Davies (ed.), Procedures in nucleic acid research, vol. 2. Harper & Row,Publishers, New York.

4. Guy,J.S.,and D. A. Brian. 1979. Bovine coronavirus genome. J. Virol. 29:293-300.

5.Hayward,W. S. 1977. Size andgeneticcontentof viral RNAs in avian oncovirus-infectedcells.J.Virol. 24:47-63.

6. Lai, M. M.C., and S. A.Stohlman.1978. RNAof mouse hepatitis virus. J. Virol. 26:236-242.

7. Lai, M. M.C.,and S. A.Stohlman.1981.Comparative analysis ofRNA genomes of mouse hepatitisviruses.J. Virol. 38:661-670.

8. Lomniezi,B., and I. Kennedy.1977. Genome of infec-tious bronchitis virus. J. Virol. 24:99-107.

9. MacNaughton, M. R., and M. H. Madge. 1978. The genome of human coronavirus strain 229E. J. Gen. Virol.39:497-504.

10.McIntosh, K. 1973. Coronaviruses: a comparative review. Curr.Top. Microbiol.Immunol.63:85-129.

11. Robb,J.A., and C. W. Bond. 1979. Pathogenicmurine coronaviruses. I.Characterization of biological behavior invitro and virus-specificintracellularRNA ofstrongly neurotropic JHM and weakly neurotropic A59V viruses. Virology94:352-370.

12. Siddell,S. G., H. Wege, A. Barthel, and V. ter Meulen. 1980.Coronavirus JHM: cell-free synthesis of structural proteinp60.J.Virol. 33:10-17.

13.Spaan, W. J. M., P. J. M. Rottier, M. C. Horzinek, and B. A. M.vanderZeijst.1981.Isolation and identifi-cation ofvirus-specificmRNAs in cells infected with mousehepatitis virus (MHV-A59). Virology 108:424-434.

14. Stern,D.F., and S. I. T.Kennedy. 1980. Coronavirus multiplication strategy. I. Identification and character-ization ofvirus-specific RNA. J. Virol. 34:665-674. 15. Stern, D.F., and S. L. T.Kennedy. 1980. Coronavirus

multiplication strategy.II.Mapping the avian infectious bronchitis virusintracellular RNAspeciestothe ge-nome. J. Virol.36:440-449.

16. Strauss, J.H., and E. G. Strauss. 1977.Togaviruses,p.

111-166.In D. P.Nayak (ed.), The molecular biology of animal viruses.MarcelDekker,Inc., New York. 17. Sturman, LS. 1977. Characterization of a coronavirus.

I. Structuralproteins: effects of preparative conditions onthe migration of protein in polyacrylamide gels. Virology 77:637-649.

18.Tyrrell,D.,J.Almeida,C.Cunningham,W.Dowdle, M.Hofstad, K.McIntosh, M. Tajima, R. Zakstel-skaya, B.Esterday, A. Kapikian, and R. Bingham. 1975.Coronaviridae. Intervirology 5:76-82.

19.Wege,H., A.Muller,andV. terMeulen.1978.Genomic RNA of the murine coronavirus JHM. J. Gen. Virol. 41:217-227.

20. Weiss, S. R., H. E. Varmus, and J. M. Bishop. 1977. The size and genetic composition of virus-specific RNAs in thecytoplasm ofcellsproducingavian sarcoma-leu-kosis viruses. Cell 12:983-992.

21.Yogo, Y., N.Hirano, S. Hino, H. Shibuta, and M. Matumoto. 1977.Polyadenylate in the virion RNA of mousehepatitisvirus. J. Biochem. 82:1103-1108.

J. VIROL.

http://jvi.asm.org/