Virus Interference by Cellular Double-Stranded Ribonucleic Acid

Copyright©1971 American Society for Microbiology PrintedinU.S.A.

Virus

Interference by Cellular

Double-Stranded

Ribonucleic Acid

P. C. KIMBALL AND P. H. DUESBERG

Department ofMolecular Biology and Virus Laboratory, University ofCalifornia, Berkeley, California 94720

Receivedforpublication8February 1971

Ribonuclease-resistant ribonucleic acid (RNA) was isolated from uridine-labeled cultures of rabbit kidney, chicken embryo, and HeLa cells. This RNA, regardless of itssource, wasfoundtoinduce interference with virus growth in either rabbit kidney or chicken embryo cultures. Nuclease-treated cellular nucleic acids exhibited interference-inducing activity which eluted witha smallfraction of RNA in the exclusion volume ofa 6% agarose gelcolumn. Besides resistance to ribo-nucleases, the interference inducer and RNA isolated from partiallydigestednucleic

acids have in common two properties of double-stranded RNA: (i) similar sharp melting profiles were obtained for inducer and ribonuclease-resistant RNA, with

TmdependentonNaCl concentration; (ii) ribonuclease-resistant inducer andRNA

banded together in Cs2SO4 density gradients at a density characteristic of known double-stranded RNA. After melting at low ionic strength, the labeled RNA shiftedto ahigher density and its capacitytoinhibit virus replication waslost.

Ve-locity sedimentation analysis of the cellular ribonuclease-resistant RNA indicated

that themajority sedimented between7and 1 S, but onlyRNAsedimentingat .8

to20Shadahigh specific activity of interference induction. Without prior

ribonu-cleasetreatment, the ribonuclease-resistant RNAcanbe precipitated with2 MLiCl

and thusappearstoexist in purified cellular nucleic acidsaspartof molecular com-plexes with both single- and double-stranded regions of RNA. Thebiosynthesis of

cellulardouble-stranded RNA is inhibited by actinomycin D.

Interferon

is

induced

by

a variety

of

viruses,

often

under

conditions

where little or no

viral

multiplication

occurs

(16).

It was

suggested,

therefore,

that some part

of

the virus particle

itself,

perhaps the

"foreign"

nucleic acid, could be

the

stimulus for induction

(18). Early

experiments

designed

to test

this

theory

by

treatment

of

a

given

species of animal

orcells

in

culture with diverse

nucleic acids

supported

the

notion

that

nucleic

acids from

heterologous

animal species could

induce

interferon (20-22,

27, 30), whereas those

from

homologous cells

were less potent or

inac-tive

(20,

22,

27) unless

chemically modified (20).

Consequently,

the

discoveries

of several

in-ducers which

apparently

contained no nucleic

acid

(15, 17)

resulted

in questioning of

thetheory

that a

foreign

nucleicacid was theessential

stimu-lus

for

interferon

induction (19). Conflicting

re-sults

in furthertests

(3, 7)

haveleft the issue about

the

"foreign" origin of inducing

ribonucleic acid

(RNA)

unresolved.

Recently,

it becameclear that

synthetic

ornatural double-strandedRNAspecies

which

arenot

endogenous

to the animal system

used

for

assay are

extremely

potent interferon inducers

(11),

whereas pure

deoxyribonucleic

acid

(DNA),

DNA-RNA

hybrids,

and

single-stranded RNA were found to be

inactive

(4,

5,15).

These

findings

and the

discovery

that

from

about

0.01 to 1.0%

of

RNA in animal cells is

ribonuclease-resistant (6,

24)

or

reactive

with

anti-bodies

specific for double-stranded

RNA

(28)

led

us to suspectthattraces

of

the

endogenous

ribo-nuclease-resistant RNA

could

be the

interferon

inducer in

cellular nucleic acids

previously

thought to

contain

RNA

only

in

single-stranded

form. Furthermore,

the

suspected

presence

of

double-stranded

RNA in normal cells raised the

question

of

whether

homologous

as well as

foreign double-stranded

RNA can induce

inter-feron in

a

given species of

cells. We present here

further

characterization

of

cellular

ribonuclease-resistantRNAandshow that it isindeed double-stranded.Part

of

thisRNAis

capable

of

inducing

interference with virus growth ina

given species of

cells, irrespective

of whether it is isolated froma

heterologous

or

homologous

species.

MATERIALS AND METHODS

Cellcultures. For

preparation

ofrabbit

kidney

pri-marycultures, mincedkidneysfrom 3-to5-week-old

New Zealand White rabbits were stirred gently at 697

on November 11, 2019 by guest

http://jvi.asm.org/

room temperature for 1 to 2 hrin a solution ofEagle's minimal essential medium (MEM; Grand Island Bio-logical Co.) containing 0.25% Pronase. The cell sus-pension resulting from one kidney pair was washed andplated in MEMcontaining 10% calf serumon a

total tissue culture dish surfacearea of about 1,000 cm2. Withamediumchange thenextday, confluency was reached in5 to 7 days after seeding. A line of HeLacells(HeLaI) from C.Colby,Jr.(Universityof

Connecticut, Storrs) and HeLa S3 cells (HeLa II) from M. Bishop (University of California, San Fran-cisco) werealso grown inMEMplus 10% calfserum.

For preparation of RNA, primary or secondary chicken embryocultures (26) weregrowninmedium

199 (GrandIslandBiological Co.) supplemented with

2%Tryptose phosphate broth and

1%

each calf and chicken sera. Periodicinspection ofallcell types by electronmicroscopy showednoevidence of contami-nation with mycoplasma.

Buffers. Low-salt buffer contained 0.01 M NaCl in additionto0.01 Mtris(hydroxymethyl)aminomethane (Tris)-hydrochloride (pH 7.4) and 1 mm ethylenedi-aminetetraacetic acid (EDTA). High-salt buffer was thesameexceptfor the NaCl concentrationwhichwas 0.3 M.

Preparation ofanuclease-resistant fraction of

cellu-lar 3H-RNA. Confluent monolayers were labeled by adding50 to 100,uCiof3H-uridine per 100-mmpetri

dish after 3to4hrofpreliminaryincubation in 4 ml of the freshgrowth mediumappropriatefor each cul-ture type as indicated above. Actinomycin D was added, whereindicated, at 6,ug/ml for 15 min before theadditionof radioactive precursor. After 90to120 min oflabeling, the nuclease-resistant RNA fraction

was prepared from purifiedcellular nucleic acids as

described earlier (6).Forsamplesofnucleic acidstobe treated with deoxyribonuclease alone, ribonuclease-free pancreatic deoxyribonuclease I (Worthington Biochemical Corp.) was used. Otherwise, the nucleic acids were incubated first with standard grade pan-creatic deoxyribonuclease Iand then with pancreatic

ribonuclease, following published procedures (6). When the partially digested cellular nucleic acids

werechromatographically analyzedon acolumn of 6% agarose gel, the exclusion volume (determined as in reference 6) routinelycontained the first two or three fractions of material labeled with 3H-uridine. These

were pooled andextractedtwice with phenolbefore precipitation with ethanol. At this stage, a visible precipitate often formed which madeaturbid solution whenredissolved in_1mlof low-saltbuffer(without

EDTA). This turbidity was largely removed from solutions without loss of radioactive RNA by centri-fugationat 150,000 X g for 10min, after which the supernatantfractionwas storedat -20C. This frac-tion is referredto hereafter as the nuclease-resistant RNAorcellular double-stranded RNA.

Standard nuclease treatments andthermal

denatura-tionofRNA.Thestandardribonucleasetreatmentwas with20,jgofpancreaticribonuclease per ml and0.5,ug

ofT, ribonuclease per ml (Worthington Biochemical Corp.) in0.8ml ofhigh-saltbuffer for30minat37C. The standard ribonuclease plus deoxyribonuclease

treatmentwasthesameexcept that50,ugof

deoxyribo-nuclease I perml and2 mmMgCl2 werealso present.

To standardize the ratio of enzyme to nucleic acid, yeast soluble RNA (sRNA; CalBiochem) wasadded (60 ,g/ml) to all samples with <0.01 absorbancy (A)260/ml. Thermal denaturation of ribonuclease-re-sistant RNA was done in low-salt buffer asreported

earlier (6).

Fractionation of RNA with 2M LiCl solution.

Purified total cellular nucleic acidswere treated with

deoxyribonuclease, extracted with phenol, ethanol-precipitated, and washed with ethanol as described

previously for preparation of the nuclease-resistant fraction of cellular RNA. The

deoxyribonuclease-treated nucleic acids were dissolved in low-salt buffer at<_10

A260/ml,

and sufficient LiCl in

solu-tion (10M) was introducedtogiveafinal

concentra-tion of2M.After 16 hr or more at0C, the solution

was centrifuged for 15 min at 20,000 X g. Three volumes oflow-saltbufferwasaddedtotheresulting

supernatantfraction and several milliliterswasadded to thepellet. Then the nucleic acids in each fraction

wereethanol-precipitated and washed.

Virusinterference assays. Assayswere done in

dupli-catebyusing0.1 ml of medium per cellculture in

6-mm wells of microtest tissue culture plates (Falcon Plastic). The assaycultures were either 7- to 14-day-old rabbit kidney cell primarycultures prepared and grown asdescribed above or chicken embryo primary

cultures plated at 5 X 104 cells per 6-mm well and incubated for 5 to 6 days in NCI Medium (Schwarz BioResearch) supplemented with 2% Tryptose phos-phatebroth and1%S eachcalf and chicken sera. Sam-ples of RNA to be tested for interference with virus replication were precipitated in ethanol with carrier yeastsRNA anddissolveddirectlyinmedium 199with

2%o fetal calf serum and 10 ,g of

diethylaminoethyl-dextranperml (molecular weight = 2 X 106, Phar-macia) ataconcentration of<100ugof RNA perml. This concentration of sRNA in 0.1 ml failedtoinduce resistance to viruses and did not affect interference

induced by the synthetic polymer poly rI:rC (Miles Laboratories, Inc.).

After 18to24hrof exposuretonucleic acid solu-tions, chicken orrabbit cells were washed with Tris-buffered saline (TS; reference 8) and challenged,

respectively, with either Sindbis virus or vesicular stomatitis virus (VSV, Indianastrain)at amultiplicity of about5plaque-forming units (PFU) per cell. Virus

was applied in 20

Muliters

of virus growth medium (NCIplus6%calfserum forSindbis, MEMplus 10% calfserumfor VSV) which was removed aftera 1 hr

period of adsorption. Then cells were washed twice withTSbeforebeingrefedwith0.2ml ofappropriate virus growth medium. After 18 to 24 hr of further incubation, cultures were examined for

cytopatho-logical effects (which correlated with viral growth), and theplates were frozenat-70 Cuntil the medium

wasanalyzedbyplaqueassayonchick embryo second-ary monolayers as previously described for Sindbis virus(6).

Control samples ofastandard interferon inducer, poly rI:rC, were included in all experiments when interfering activity of cellular nucleic acids was as-sayed.PolyrI:rCreducedvirus growth to titers below

on November 11, 2019 by guest

http://jvi.asm.org/

detectionataconcentration of 0.1

Ag/ml

orlesswhen

cultures without inducergave106to 107 PFU/ml for

either the chicken orrabbit cell system. Results are

reportedasthelog10(tothenearest0.1)of theratio of virustiter inuninduced control culturestotheaverage

titer intwoparallel culturestreatedwithagiven

sam-pletobe assayed.This ratio isreferredtoaslog10virus

titer reduction, andin practice duplicateassays gave

reductions differing by _0.5 log10 Since, ingeneral,

the log1o virus titer reduction was not found to be linear with inducer RNA concentration, no attempt

wasmadetoquantify differencesinRNA concentra-tion of compared samples when they caused titer

reductions which varied outside the range of

dupli-cates.Weused this titer-reductionassay as ameasure ofinterferon induction because it has been shownto be more sensitive than assays for extracellular

inter-feron (5, 31), even before adaptation to microassay procedure.

RESULTS

Isolation ofaribonuclease-resistant interference inducerfromcellular nucleic acids. To answer the general question of whether the cellular nucleic

acids ofa given species contain double-stranded

RNA abletoinduceinterferonin thesamespecies

orotherspeciesofcells,thefollowing experiments

werecarriedout.Thenuclease-resistantRNAwas

preparedfrompurifiedtotal cellular nucleic acids of chicken embryo fibroblasts or rabbit kidney

cells by treatment with deoxyribonuclease and

ribonuclease and chromatographic fractionation

of thedigestasoutlinedabove.Onlythatfraction

of the ribonuclease-resistant RNA which eluted in the void volume of the 6% agarose column was

thenanalyzedfor itsabilitytointerfere with virus

replication. In all four tested cell types, about 0.4% of the total 3H-uridine incorporated into RNA during a 90-min pulse was in the

ribonu-clease-resistant fraction excluded by 6% agarose

(Table 1).Thisamounted to less than 0.1% of the totalA26oin thedigest.

Nuclease-resistant RNA from chicken, rabbit,

or HeLa cells depressed virus growth in both rabbitkidneyand chickenembryocell

cultures-in all cases to a level lower than 2 log10 below

control titers. By contrast,the same amounts of

totalcell nucleic acids whichwereused toisolate

the interference-inducing, ribonuclease-resistant

RNA did not showanyinterfering activitybefore

digestion with deoxyribonuclease and ribonu-clease. After nuclease treatment, however, samples of unfractionated nucleic acids from all cells interferedwith virusreplicationtothesame extent (within 0.5 log10 virus titer reduction) as

thepurifiedribonuclease-resistant fraction alone. These findings are summarized in Table 1. De-oxyribonuclease treatment alone also rendered cellular nucleic acids active ininducing

interfer-ence with virus

replication (data

not shown). Furthermore, insome

experiments,

partof the cell

suspension

from either one chicken embryo or

rabbitkidneywasculturedfor interference assays, whereas the rest of the cells were grown as a source of labeled

ribonuclease-resistant

RNA. Thus induction of virus interference by the pre-sumeddouble-stranded RNA

fraction of

chicken

or rabbit cell nucleic acidswasdemonstrablenot

only

in cells of the sameor different

species, but

also

incells derived from thesame

piece of

tissue

as

those

from which the

ribonuclease-resistant

RNAwas

prepared.

Further

nuclease

treatments

of

the 3H-RNA

purified from

degraded

cellular RNA and DNA

by

chromatography

on

6%o

agarose were

per-formed

under the standard conditions

(see

above)

at

high ionic strength.

The

resistance

to

ribo-nuclease

(or

amixture

of ribonuclease

plus

deoxy-ribonuclease)

of this fraction of

3H-RNA ranged from 50% togreaterthan

95%0

in different

prep-arations from

eachcell type. However, no

signifi-cant change

(.0.5 log10

virus

titer

reduction)

in

the

ability

of

these RNA

fractions

to

interfere

with

virus

replication

was

observed

after standard

nucleasetreatments

of

seven

of

eight

RNA

sam-ples

from chicken and

HeLa cells

(data

not

shown).

After thermal

denaturation

(Materials

and

Methods) of

the

purified

ribonuclease-resis-tantfraction

of

cellular RNAat 100

C,

less than

5% of

the 3H-RNA remained

resistant

to the

standard

ribonuclease

treatment and all

interfer-ing activity

waslost

(Table 1).

The

finding

that the

nuclease-resistant

fraction

of cellular

RNA is resistant to both

deoxyribo-nuclease and

ribonuclease

argues

against

the

possibility

that this

ribonuclease-resistant

3H-RNAwhich interferes with virus

replication is in

DNA-RNA

hybrids

because DNA-RNA

hybrids

are

known

to be

sensitive

to

deoxyribonuclease

(25)

or a

combined

nuclease treatment

(32).

However,

double-stranded

RNA

would

be

ex-pected

both to resist ribonuclease at

high ionic

strength

and to be

made

sensitive

by

melting at

low

ionic

strength

before ribonuclease

treatment

(14).

In

addition, only

such RNA is known to

showa

high

specific activity of

induction of

virus

interference (4,

5,

15).

It can be estimated from

thedata

in

Table1that less than 0.05

,ug (

<0.01

A26o/ml)

of ribonuclease-resistant RNA per ml reduced titers oftwodifferent virusesbyasmuch

as 5

log10

in two different cell

species.

Together

these observations suggest that the ribonuclease-resistant

fraction of

cellularRNAwhich

interferes

with virusgrowth is double-stranded RNA. Interference-inducing RNA appears to be

par-tially

single-stranded before ribonuclease

treat-ment. To determine whether the RNA in the

7,

on November 11, 2019 by guest

http://jvi.asm.org/

[image:4.493.57.455.88.425.2]

TABLE 1. Initerference with virusgrowthby cellular nucleic acids before antd after ntuclease treatments

Cellspecies

Rabbit kidney

Nucleic acid fraction

Total purified nucleic acids

Nuclease-resistant frac-tion excluded by6%c agarose

Chicken Total purified nucleic

embryo acids

Nuclease-resistant frac-tion excluded by6%C agarose

Totalpurified nucleic acids

Nuclease-resistant frac-tion excluded by6%

agarose

Total purified nucleic

acids

Nuclease-resistant frac-tion excluded by6% agarose

Trichloroacetic acid-precipitable 3H-nucleicacids Nuclease treatmenta

None

Deoxyribonuclease,

thenribonuclease None

Melted, then standard ribonuclease treat-ment

None

Deoxyribonuclease,

then ribonuclease

None

Melted, then standard ribonuclease treat-ment

None

Deoxyribonuclease,

thenribonuclease

None

Melted,thenstandard ribonuclease treat-ment

None None

Countsper Per

centi

min perml of total

600,000 100

2,700

<10b

0.45

845,000

1100

2,600

<110

0.31

510,000 1100 1,170

<50

580,000 2,500

0.23

100

0.431 <0.001

aDeoxyribonuclease and ribonuclease treatments of the total nucleic acids are described for the

preparation of anuclease-resistant fraction of cellular 3H-RNA.Melting was performed at 100 C in low-salt buffer. Theprocedureis presented in Materials and Methodswiththe details of the standard ribonucleasetreatment.

bNotcorrectedforbackground.

cEstimated amount ina10-folddilution of theoriginalsolution which read<0.01A260/ml.

dDefined in Materials andMethodsunder virus interferenceassays.

nuclease-resistant fraction of cellular nucleic acids

wascompletely oronly partiallydouble-stranded

before ribonucleasedigestion,itssolubilityin 2M

LiCl was examined. It is known that

double-stranded RNA is soluble but partially double-stranded RNA is insoluble in 2M LiCl (1). After

deoxyribonucleasetreatment,total cellular nucleic acids werefractionated with 2 M LiCl according

totheproceduredescribedabove. Then the super-natant and pelleted nucleic acids were either

assayed directly for virus-interfering activity or digested withribonuclease before isolation of the nuclease-resistant fraction of 3H-RNAbyagarose

column chromatography (see above). Over 90%0

of the total ribonuclease-resistant 3H-RNA in

HeLa Iorchicken embryo cell nucleic acidswas

found in the LiCl-insoluble pellet. Moreover, interference-inducing activitywasdetectable only

in the precipitated fraction ofHeLa I or rabbit kidney cell nucleic acids. These results are

sum-marized in Table 2. Under thesame conditions, only

12%7

(195 of 1,560 counts/min) of added influenza virus double-stranded 32P-RNA (10) was precipitated with HeLa I cellular

double-stranded RNA. Therefore, precipitation in 2 M

LiClof theinterference-inducing activity withthe

ribonuclease-resistant RNA suggests that, in the total purified nucleic acids of the cell, the nuclease-resistant inducer could be the

double-stranded coreofan RNAcomplex thatalso has Interference assays(logiovirus

titerreduction)d

A260/ml

4.0

<O.OOlc

3.5

<0.001

1.1

<0.001

3.6 HeLa I

HeLa II

Chicken cells

-0.2

2.4

0.0

3.9

3.8 0.2

0.1

4.2

-0.3

3.4

Rabbit cells

0.0 4.0

4.5 -0.2

-0.3 3.6

3.3

-0.2

0.0

4.3 4.3

0.0

5.3

on November 11, 2019 by guest

http://jvi.asm.org/

TABLE2. Distribution in 2 MLiCI of ribonuclease-resistant 3H-RNA and virus-intterfering activity

of cellular nucleic acidsa

2MLiCI-soluble 2MLiCI-insouble fraction fraction

Cellular Ribo- Viu nVirus

in-nucleicacids nuclease- Virus Ribonuclease-

terferencec

resistant ererence resistant (oi

(logA

s RNAb (logirstie

(counts/ virustiter (counts/min.,) rustiter

min) reduction) reduction)

HeLa I 125 (9)d -0.2

11,240

(91)d 3.7

Chicken

embryo 77 (10) 785 (90)

Rabbit

1-kidney 0.1 4.4

aTotal purified nucleic acids were treated with deoxyribonuclease and then fractionated with 2 M LiCI.

bThe 2 M LiCi-soluble and -insoluble nucleic acids were treated with ribonuclease, and the ribonuclease-resistant fraction of RNA was puri-fied by agarose gel exclusion chromatography as described for preparation of the nuclease-resistant fraction of total cellular nucleic acids.

c After precipitation with ethanol, the nucleic

acids in theLiCl-soluble and -insoluble fractions wereassayed for interference with thereplication ofvesicular stomatitis virus on rabbit kidneycells.

dNumbers in parentheses are expressed as per cent.

single-stranded

regions

whichcause

precipitation

at

high ionic strength.

Melting

profiles of

nuclease-resistant

RNA and

interference inducer. As

expected for

an RNA

duplex

(20),

melting

profiles with sharp

transi-tions

(at

86 C in 0.01 M

NaCI)

from ribonuclease

resistance

to

ribonuclease

sensitivity

were

ob-tained

for the

purified

nuclease-resistant

fractions

of the different

cellular

RNA

species (Fig. 1A,

B and

C).

In the case

of

HeLa I

nuclease-resistant

RNA, the

ability

to induce virus interference in rabbit

kidney

cultures was

completely

lost when

the

3H-RNA

was

completely

melted

(Fig. IC).

Measurement of the

T,,,

of cellular

double-stranded RNAon the basis of its

interfering

ac-tivity

in thevirus titer

reduction

assay

would

re-quire both

the measurement

of

a

twofold

(0.3

logls)

virus titer

reduction

and the

assumption of

someprecise relationship between the

concentra-tion of

interference-inducing

double-stranded

RNA and titer reduction. Our

experiments

were

not

designed

to determine either such a small

change in virus titer or the exact nature of the

relation between

inducing

RNA concentration and titerreduction

(see

above forthelimitsof the interference

assay).

To test whether the

T,,,

of the material which

interferes with virus growth is increased with in-creasing ionic strength as has been shown for double-stranded RNA (2), HeLaI ribonuclease-resistant 3H-RNA was heated at 100 C in high-salt buffer with 0.3 M NaCl (see above). It was

found that 21

%/o

(95counts/min) of the 3H-RNA and enough virus-interfering activity to give 1.5

log1o

virus titer reduction remained resistant to

ribonuclease after heating under these conditions. By contrast, less than 3% of the 3H-RNAand no measurable interfering activity (-0.1 log10 virus titer reduction) survived the standard ribonu-clease treatment after heating to 100 C at low ionic strength (Fig. IC). This indicates that the

T,-

for ribonuclease resistanceof the interference-inducing 3H-RNA was increased with increasing ionic strength.

Buoyant density in CS2SO4 of ribonuclease-resistant RNA and inducer of virus interference beforeandafter

melting.

To examine further the

suggested identity

between the

purified

ribonu-clease-resistant interference inducer and double-stranded RNA, the density distributions of 3H-RNAand inducerweredetermined

in gradients

of Cs2SO4. In all cases, the

ribonuclease-resistant

3H-RNA banded in a sharp peak at adensity of 1.61 to 1.63

g/ml.

Single-stranded tobacco

mosaic virus (TMV) RNA marker hada

density

of 1.68 g/ml, and salmon sperm DNA had a

density

of less than 1.50 g/ml under the same conditions. Interference assays on rabbit

kidney

cells showed

higher

levels of antiviral

activity only in

the

gra-dient fractions

containing

the

peak of

the

nuclease-resistant

3H-RNA

(Fig. 2A,

B,

and

C).

The

density of

the

biologically active

RNA

frac-tion

is in

good

agreement with that

of known

double-stranded

RNA

(29)

and is much greater

than those

reported

for DNA-RNA hybrids (1.49 to 1.54 g/ml;

references

9, 29). When a

sample of ribonuclease-resistant

3H-RNA

from

eachcell typewasmeltedat100Cbefore

analysis,

the 3H-RNA and TMV RNA marker

sub-sequently

banded

together

at 1.68

g/ml

(Fig.

2D,

E,andF), and

interference-inducing capacity

was

lost.

Similar

density

shifts and loss

of

activity

were

noted after treatment of ribonuclease-resistant

RNA with

97%NO

dimethylsulfoxide (data

not

shown).

The

temperature-dependent

density

tran-sition of

the

ribonuclease-resistant

3H-RNA is

consistent with the melting of an RNA

duplex

(29),asistheloss ofability to induce virus inter-ference

(6).

Sedimentation

properties

of the interference-inducing nuclease-resistantRNA before and after

standard nuclease treatments. The ribonuclease-resistant 3H-RNA

separated

from cellular nucleic

acid

digests

by agarose

gel chromatography

sedi-mentedthroughsucrose

gradients

inabroadband

on November 11, 2019 by guest

http://jvi.asm.org/

[image:5.493.44.236.76.243.2]

CHICK EMBRYO

-Tm =86°C

40 50 60 70 80 90 100 40 50 60 70 80 90 100

[image:6.493.63.455.58.310.2]

T E M PER ATUR E C

FIG. 1. Meltinigprofiles of niuclease-resistanit RNAantd virus interference iniducerpurified from ntucleic acid

digests byexclusionchromatographyoni6%agarose. Samples ofRNA wereheatedattheindicatedtemperatures anidassayedfortrichloroaceticacid-precipitable radioactivity afterthe standardribonucleasetreatment.ForpartC, samples of niuclease-resistanit HeLa cell RNA were also analyzed for interference inducer. After thermal de-niaturation andtreatment with ribon2uclease unider thestandardconditions, theRNA wasphenolextractedthree

timesbeforeprecipitationi with carrier sRNAbyadditionoftwovolumesofethanol. TheRNAwasthenidissolved

directly inthe mediumu*sedjbrviruisiniterferenceassayswithrabbitkidnzeycellcultures. with a major peak occurring at about 7S for

rabbitkidney, 9Sforchickenembryo,and11S for HeLa I cells whencomparedwith4S yeast sRNA marker(Fig. 3).Sedimentation coefficientsranged from6to8S forrabbit cellribonuclease-resistant

RNA (seven gradients of fourindependent

prep-arations), 7 to 9S for chicken cell RNA (six gradients offour preparations), and from 10 to 11S for HeLa I cell ribonuclease-resistant RNA

(four gradients oftwopreparations). Foragiven preparation, little change in sedimentation be-havior of the ribonuclease-resistant 3H-RNAwas

observed after the standard nuclease treatments,

althoughstillmoreribonucleasedigestionreduced

the averageS valuebymorethan 15 S (datanot shown). Similar results were reported for de-pendenceofsedimentationon ribonuclease treat-ment in the case of MS2 double-stranded RNA

(2).

The lowest sedimentation coefficient obtained inourexperiments after the standard deoxyribo-nuclease plus ribonuclease treatment of purified ribonuclease-resistant RNA ispresentedforeach cell species in Fig. 4. A shoulder (Fig. 3B) or

distinct secondcomponent (Fig.4B)with alower

sedimentationcoefficient than the mainpeakwas

observed in most preparations of chicken cell

RNA. We did not examine the possibility that

the physiological state of the cells in culture

determines the sedimentation characteristics of

different isolates of cellular double-stranded RNA. Nor was it shown whether the variations

reflect differences in the original size or in the

ribonuclease sensitivityofthebase-paired portion

oftheundigestedmolecules. In allcases,however, it wasfoundthat the purified ribonuclease-resis-tant RNA which induced virus interference was

distributed insucrosegradientsfromover7Sto at least20S. Most of the active fractions (15to40% of thetotal) weresedimented faster than the main peaksofcellularnuclease-resistant 3H-RNA (Fig. 4).

Usingthe formula of Franklin for the

molecu-lar weight of double-stranded RNA (molecular weight = 2.42S; 16; reference 13), one can

esti-mate average molecular weights for cellular double-strandedRNA of 5 X 104daltons for 7S rabbit kidney cell RNA,2 X 105for9S chicken cellRNA,and 6 X 105 for11S HeLa I RNA. The

interference-inducing double-stranded RNA, with

sedimentation coefficients from 8 to 20S, canbe estimatedtopossess aminimum molecularweight

of 105daltons andamaximum of several millions.

A minimum molecular weight of 1.5 X 105 to

B c

HHE LAI Tm =86°C

*

0

I

0

!

l

E

UL

3.0

--z

0 0 2.0 a 0

U

cr-1.0 e)

o

0

0 0.0

Ij

r-L-i

L

on November 11, 2019 by guest

http://jvi.asm.org/

RNA

703

1.8

1.7

1.6

1.5 1.4 1.3

1.8

-1.7 -1.6

-1.5

-1.4

0

E

"I

en z 0 z

0

0

()

JI.3

1.5

1.0

0

0

cD

cx

I

5.0 ₀Z

H

4.0 2

0

3.0 n 2.0 f 1.0 ci

0.0

>

-0.1 £

0

0

-J 1.5

1.0

0.5

0

FRACTION NUMBER

FIG. 2. Density distributionz of cellular nuclease-resistant 3H-RNA anid virus initerferenice in2ducerin Cs2SO4

gradienztsbefore antdafterthermal deniaturation. ForpartsA, B,andC, samples ofinticlease-resistantRNAfractions

wereanalyzed afterafurther stantdardnucleasetreatment,followed bythreepheniolextractions anidethaniol

pre-cipitationiinlthepresence of onteA260 unit eachoftobaccomosaicvirusRNAandsalmoniDNA. ForD, E,andF, sampleswereheated inlow-saltbufferat100 Cfor3miti,quenchedinmeltinigice,andprecipitatedwithmarker

niucleicacidsasabove. Eachsample wasthendissolved in 3.8 ml ofTris-hydrochloride buffer (0.05 mi,pH 7.4) containinlgsuifficient Cs2S04togiveafinalaveragedensity ofabout 1.58g/ml. Centrifugation wasat20 Cantd

31,500rev/mi,zfor3daysinaSpincoSW 50.1 rotor.Gradientswerefractioniated bycollecting dropsfroma

pin-hole in the bottomof the centrifuge tubes.Samples of gradient fractionswere takentup immediatelyin0.100-ml

taredmicropipettes whichwereweighedtomeasuresolutiondensity. Toeachfractioni anequal volume ofwater wasaddedbefore readingA2M0,andsampleswereassayedeitherfortrichloroacetic acid-insolubleradioactivityor,

afterethantolprecipitation, for initerference inlducerinrabbit cells. The latterassays were nlotperformed forparts D, E, anid F becausemelting destroys the inducer. Theslight interference inductioniobservedunderthe tobacco mosaic virus RNAmarkerforthegradientinpart Bprobablyindicates nuclease-resistantRNAco-precipitating

withsinigle-strandedRNA duetoincompleteremovalofsingle-stranded regions fromthenuclease-resistantcoreof

theRNAcomplex.

1.6 X 105 daltons for interferon induction has been determined similarly for the synthetic polymer duplexrI:rC (12).

Effects of actinomycin D onthe biosynthesisof cellular double-stranded RNA. Incorporation of 3H-uridine into the double-stranded RNA frac-tion of cellular nucleic acids was inhibited by actinomycin D, as reported previously (24). In-hibitationwas 90%c ormoreinthetwospeciesof cellsgrowninprimary cultures, whereas the

syn-thesis of ribonuclease-resistant 3H-RNA in the

two HeLa cell lines was reduced by 78 to 85% (Table 3).Totalincorporationof3H-uridine into cellular RNA was inhibitedby over 98% under

thesameconditions(datanotshown).The inhibi-tion of incorporation into cellular double-stranded RNAbyactinomycinDsuggeststhatits

synthesismaybedependentonDNAtemplate. DISCUSSION

Experiments described in this paper indicate

that animal cellscontain afraction of RNA that

0

E

I

on November 11, 2019 by guest

http://jvi.asm.org/

[image:7.493.45.442.58.384.2]

500 II I

A 7S 4S B C iis 4S

RABBIT CHICK EMBRYO HELA I

KIDNEY 400

~~~~~~9545

_ 300 3.0

E

0.

0

4~~~~~~~~~~~~~~~~~~~~-.

-200

2.00

100 1.0

0 5 10 15 20 0 5 10 15 20 0 5 10 15 20

FRACTION NUMBER

FIG. 3. Sedimentationanalysis of purified cellular ribonuclease-resistant RNA. Samples were precipitated wit/l twovolumesof ethanol in the presence of 50

lug

of 4S yeast sRNA marker. Pellets were then dissolved in 0.15 ml of buffer containing0.01 M Tris-hydrochloride (pH 7.4), 0.1 m LiCl, Imm EDTA, and

0.1%1,

sodium dodecyl sulfate and layered on 5-mllineargradienitsof sucrose (10 to25%) in the samebuffer.A0.3-ml cushion of65% sucrose in D20 was included at the bottom of each gradient. Centrifugation was for 315 min at 20 C and 65,000 rev/min in a SpincoSW 65rotor.Gradients were fractionated as in Fig.2and weresimilarly assayedforA260 and trichloroaceticacid-precipitable radioactivity.

A,I II,

RABBIT KIDNEY dsRNA CHICK EMBRYO dSRNA HELA I dsRNA

75 4S 7S 4S 7S 4S

700 -SS7

600 5.

1

I A X0oJo

0.5 -4.0

400

10015

20 0 5 10 15451 2 -3.0

300

-~~~~~~~~~~~~O3 2.0wu

0

] H

I

0.2 1.0 U

100

~~~~~~~~~~~~~~~~~~0.1

-0.0

>

0

0 5 10 15 20 0 5 10 15 20 0 5 10 IS 20

-Ol

FR ACT I ON NU M BE R

FIG.4. Sedimentation distribution of interference-inducing activityandradioactivityafternuclease treatment of cellular double-strandedRNAlabeled with 3H-uridine. Aftertheribonuclease-resistant RNAfractionwasgiven the standard ribonucleaseplus deoxyribonuclease treatment, it wasphenol-extractedandprecipitatedwithethanol. Velocity sedimentationanalysison sucrosegradientsisdescribedinFig.3. In thiscase,afterA260readings, partof eachfractionwasassayedfor trichloroacetic acid-insolubleradioactivity, and the remainderwasethanol-precipitated forvirusinterferenceassayswith rabbitkidneycultures.

on November 11, 2019 by guest

http://jvi.asm.org/

[image:8.493.62.461.70.321.2] [image:8.493.68.454.400.595.2]

TABLE 3. Effects ofactinomycin D on labeling of

riboniuclease-resistant sH-RNAa

Trichloroacetic-acid-precipitable

3H-RNA

(counts/min) Per cent

Cell species inhibition of

labeling Without With

actinomy-cinD cin D

Chicken

embryo

11,200

557 95

Rabbit

kidney

927 95 90

HeLa. 5,018 1,095 78

HeLaII... 1,639 242 85

aNuclease-resistant fraction of cellular

3H-nucleic acids was prepared from parallel cultures with and without actinomycin D.

is

double-stranded

and induces interference with

virus

replication.

In this study, it has not been

demonstrated

directly

that interference with the

growth

of

Sindbis

virus or VSV by cellular

ribo-nuclease-resistant

RNA ismediated by interferon.

Butexperiments carried out under similar

condi-tions

in

this

(5, 6)

and

other

(12,

13)

laboratories

have

suggested

that

induction

of interferon by

double-stranded

RNA

is the

basis of such

inter-ference.

Cellular double-stranded RNA

and

interfer-ence-inducing

activity purified from cellular

nucleic

acids share

these

properties: they are both

resistant

to

ribonuclease and

deoxyribonuclease,

they

have

the

same

distinctive

solubilities

in

salt

solutions

of

different concentrations,

and the Tm

values

for

their

thermal denaturation

are

co-ordinately affected by changes in ionic

environ-ment.

Also

they have the same buoyantdensity in

Cs2SO4.

However,

the

distributions

of

ribonu-clease-resistant interference-inducing activity

and

RNA were

only partially coincident after

frac-tionation

by

velocity

sedimentation: inducing

RNA

sedimented

faster

than the average-sized

molecule of cellular double-stranded

RNA.

Therefore,

it

may be

concluded

that the

high-molecular-weight double-stranded

RNA in

cells

is

the most

efficient

component of

cellular

nucleic

acids reported

to induceinterferon.

Earlier

reports on

ribonuclease-resistant

cellu-lar

RNA

did

not show

induction

ofvirus

inter-ference(6,

24),

but

presumably

thenegative result

(6)

was due to lower

sensitivity of

theassay sys-tem which

employed cultures

100 times

larger

than those

used

here. The

observation

inthis work

that

total cellular nucleic

acids had no

interfer-ence-inducing

action but were rendered potent

after

deoxyribonuclease

treatment cannot be

ex-plained

from our

experiments.

This work shows that

cells of

a

given animal

produce double-stranded RNA which induces

virus interference

inother cells of the same

ani-mal, as well as in those of another species of animal. This evidence indicates that double-strandedness may be the onlyessential property

for interferon

induction by

natural

RNA, rather

than the source of the RNA as postulated pre-viously

(18).

Such aconclusion raises the question of whyresident cellular double-stranded RNA, as

it

exists in the cell, does not induce detectable interference with virus growth. Perhaps the

in-tracellular concentration

of

double-stranded

RNA

is

too low for interferon induction. From

long-term

labeling experiments

with

32P-phosphate, it

is

possible

to

estimate

that of the

total

RNA

(10-

jig

per

chicken cell) only about

0.01% (~-10" Mg)

is

double-stranded

(our

unpublished

data),

in

agree-ment

with

immunochemical

measurements

(28).

The minimumlevel of poly

rI

:rC forinterference induction was found to be 10-8

Mig

per chicken

cell

(5). Assuming

that

poly

rI:rC

and

cellular

dou-ble-stranded

RNA have

similar specific

activities

of

interferon

induction,then it may be

estimated

that about

10

times

more

extracellular

double-stranded

RNA than

is

present in the

cell

is

necessary for induction of virus growth

interfer-ence.

It

is

possible, however,

that

enough

double-stranded RNA is present in the

cell

to cause

inter-ference

with viruses, but the cellular

double-stranded

RNA in

situ

may not

be able

to

interact

with

components necessary for

interferon

induc-tion.

It is

also

possible

that in the

cell less

RNA

exists in the

double-stranded

form before

de-proteinization

which occurs

during isolation.

The

biological

role

and

origin of the cellular

double-stranded

RNA are

still

open to

specula-tion: this

RNAmay be

generated by

alatent virus

infection

or

it

could be

atrue

cellular

component.

The

general

occurrence

of

double-stranded

RNA

in different

primary cells (chicken,

rabbit, mouse)

and

cell lines (human, hamster;

compare

this

report,24,

28,

and our

unpublished observations)

argues

against

a

viral

origin of this

RNA. More-over,

the

finding

that the

biosynthesis

of

cellular

double-stranded

RNAis

inhibited

by

actinomycin

Dfurther

dispels

the

suspicion

that

this

RNAmay represent the

replicative

intermediate of

a

latent

RNA

virus,

because

viral

RNA-dependent

RNA

synthesis

is

generally

known to be resistant to

actinomycin

D. However, we cannot

exclude

the

possibility

that the double-stranded RNA

might

derive from infection

by

aDNAvirus

which,

like vaccinia virus

(6),

is

capable

of

synthesizing

double-stranded

RNA.

It seems more

likely,

however,

that the

inter-ference-inducing,

double-stranded

RNA

de-scribed

here is

transcribed

from

complementary

7,

705

on November 11, 2019 by guest

http://jvi.asm.org/

[image:9.493.46.237.75.215.2]

regions ofcellular DNA. This idea is compatible with the ubiquity ofthis RNA speciesin animal

cellsandwiththesensitivity of itsbiosynthesisto

actinomycin D. It is also conceivable that some

cellular RNA-dependent RNA synthesis exists,

since we found that the biosynthesis of cellular double-stranded RNAappearsto be less sensitive toinhibition by actinomycinD than total RNA synthesis (alsomentioned in24).

Furthercharacterization of the cellular

double-strandedRNA will be necessary to speculate on

its biological role.

ACKNOWLEDGMENTS

We thank Harry Rubin forgenerouscooperation andG. S. Martin,M.Bissel,and Bart Sefton for useful criticism of the

manu-script.Furthermoreweappreciatethepleasant exchangeofresults

before publication with Edward De Maeyer, Jacqeuline De Maeyer-Guignard, and Luc Montagnier, who haveshown that

the double-strandedRNA from rat cellsinduces interferon

pro-duction incellsof thesamespecies.

This investigation was supported byPublic Health Service

grantCA 11426 fromthe National Cancer Institute andtraining

grantGM 01389 from the National Institute of General Medical

Sciences.

ADDENDUM IN PROOF

Results similar tothosereported byus have been

published byE. DeMaeyer,J.DeMaeyer-Guignard,

and L. Montagnier [Nature New Biology (London)

229:109-110, 1971].

LITERATURE CITED

1. Baltimore,D. 1966.Purification and properties of poliovirus

double-strandedribonucleicacid. J. Mol. Biol.18:421-428.

3. Billeter, M. A., C. Weissmann, and R. C. Warner. 1966. Replication of viral ribonucleic acid. IX. Properties of double-stranded RNAfromEscherichia coli infected with bacteriophageMS2.J. Mol. Biol.17:145-173.

3. Billiau, A., andE.Schonne. 1970. Inductionof the interferon mechanism by natural RNA.LifeSciences 9:69-78.

4. Colby,C. 1970. The induction ofinterferonby natural and

synthetic polynucleotides, p. 1-32. In W. Cohn and N.

Davidson (ed.), Progress in nucleic acid research. Aca-demic Press Inc.,NewYork.

5.Colby, C., and M. J. Chamberlin. 1969. The specificity of interferon induction in chick embryo cells by helical RNA. Proc. Nat. Acad. Sci. U.S.A. 63:160-167.

6.Colby, C.,and P. H.Duesberg.1969. Double-stranded RNA invaccinia virus infected cells. Nature (London) 222:940-947.

7. Colobert, L.,and P. Louisot. 1968. Inhibition of the

multipli-cation ofmyxovirusand arbovirusby chemicallymodified

ribonucleic acids from the host cells. Biochem. Biophys.

Res. Commun.30:148-152.

8. Duesberg, P. H. 1969. Distinct subunitsof the ribonucleo-proteinof influenzavirus. J. Mol. Biol. 42:485-499. 9. Duesberg,P.H., and E. Canaani. 1970. Complementarity

be-tweenRoussarcoma virus (RSV) RNA andthe in vitro-synthesizedDNAof thevirus-associated DNApolymerase. Virology 42:783-788.

10. Duesberg, P. H., and W.S.Robinson. 1967.On thestructure

andreplication of influenza virus. J. Mol. Biol. 25:383-405.

11. Field, A. K., A. A. Tytell, G. P. Lampson, and M. R.

Hilleman. 1967. Inducers of interferon and host resistance. II. Multistranded synthetic polynucleotide complexes. Proc. Nat. Acad. Sci. U.S.A. 58:1004-1010.

12. Field, A. K., A.A.Tytell,G. P.Lampson, M. M. Nemes, and M. R. Hilleman. 1970. Double-stranded polynucleo-tidesasinterferoninducers.J. Gen.Physiol.56:90S-96S. 13. Franklin, R. M. 1967. Replication of bacteriophage

ribo-nucleicacid:somephysical properties of single-stranded, double-stranded, and branched viral ribonucleic acid. J. Virol. 1:64-75.

14. Geiduschek, P. E., J. W. Moohr, andS. B. Weiss. 1962. The secondarystructure ofcomplementary RNA. Proc. Nat. Acad. Sci. U.S.A. 48:1078-1086.

15. Hilleman, M. R. 1968. Interferon induction and utilization.

J.Cell.Physiol. 71:43-60.

16. Ho, M. 1966. The production of interferons, p. 21-52. In N. B.Finter (ed.), Interferons. North-Holland Publishing Co., Amsterdam.

17. Ho, M., K. H. Fantes, D. C. Burke, and N. B. Finter. 1966. Interferons or interferon-like inhibitorsinduced by

non-viral substances, p. 181-201. In N. B. Finter (ed.), Inter-ferons.North-HollandPublishing Co., Amsterdam. 18. Isaacs, A. 1961. Mechanisms of virus infections. Nature

(London) 192:1274.

19. Isaacs,A. 1965. Studies on interferon. Aust. J. Exp. Biol.

Med. Sci. 43:405-412.

20. Isaacs, A.,R. A.Cox, andZ.Rotem.1963. Foreign nucleic

acidsasthe stimulus tomakeinterferon. Lancet2:113-116.

21. Jensen, K. E.,A. L.Neal, R. E. Owens, and J. Warren. 1963. Interferonresponsesof chickembryo fibroblaststonucleic

acids and relatedcompounds.Nature (London)

200:433-434.

22. Kohlhage, H., and D. Falke. 1964. Vermehrungshemmung des herpes-simplex-virus durch ribonukleinsaiuren. Arch.

Gesamte Virusforsch.14:404-409.

23. Louisot, P., L. Colobert,C. Bollack,J. H. Weil, and J. P. Ebel. 1968. Inhibition dela multiplication viraleal'aide

d'acides ribonucleiques chimiquement modifies.I.

Inhibi-tion de la multiplication de myxo',irusparainifluienzae I (virus sendai) surcellules de rein deveau.Biochim. Bio-phys. Acta 155:38-50.

24. Montagnier, L. M. 1968.Presence d'un acideribonucleique

endouble chaine dan descellulesanimales. C. R. Acad.

Sci.Paris 267:1417-1420.

25. Nygaard,A.P., andB.D. Hall.1964. Formation and

proper-ties of RNA-DNAcomplexes.J.Mol. Biol. 9:125-142. 26. Rein, A.,and H. Rubin. 1968.Effects of localcell

concentra-tionsuponthegrowthof chickembryo cellsintissue

cul-ture.Exp. Cell Res. 49:666-678.

27. Rotem, Z.,R. A.Cox, andA.Isaacs.1963.Inhibition ofvirus multiplication by foreign nucleic acid. Nature (London) 197:564-566.

28. Stollar, V., andB. D.Stollar. 1970.Immunochemical

meas-urementof double-stranded RNA of uninfectedand arbo-virus-infected mammalian cells. Proc. Nat. Acad Sci.

U.S.A. 65:993-1000.

29. Szybalski, W. 1968. Use of cesium sulfate forequilibrium density gradient centrifugation,p.330-360.InL. Grossman

andK.Moldane (ed.), Methodsinenzymology, vol. 12B. Academic Press Inc., New York.

30.Takano, K.,J.Warren,K. E.Jensen,and A. L.Neal. 1965.

Nucleic acid-induced resistance to viral infection. J.

Bac-teriol. 90:1542-1547.

31. Vilcek, J.,M. H.Ng,A.E.Friedman-Kien, andT.Krawciw.

1968. Induction of intetferonsynthesis by synthetic

double-strandedpolynucleotides.J. Virol. 2:648-650.

32.Yankofsky,S. A.,and S.Spiegelman.1962. Theidentification

ofthe ribosomal RNA cistron by sequence

complemen-tarity. I. Specificity of complex formation. Proc. Nat.

Acad. Sci.U.S.A. 48:1069-1978.

on November 11, 2019 by guest

http://jvi.asm.org/