Properties and Location of Poly(A) in Rous Sarcoma Virus RNA

Copyright0 1974 American Society forMicrobiology Printed in U.S.A.Vol.14,

Properties and Location

of

Poly(A)

in

Rous

Sarcoma

Virus

RNA

LU-HAI WANG AND PETER DUESBERG

Department of Molecular Biology and Virus Laboratory, UniversityofCalifornia,Berkeley,California 94720 Received forpublication28 August 1974

The poly(A) sequence of 30 to 40S Rous sarcoma virus RNA, prepared by

digestion oftheRNA with RNase T, showed a ratherhomogenous

electropho-reticdistributioninformamide-polyacrylamide gels. Its size was estimated tobe

about 200 AMP residues. The poly(A) appears to be located at or near

the 3' end ofthe 30 to 40S RNAbecause: (i) it containedone adenosineper 180

AMP residues, and because (ii) incubation of 30 to 40S RNA with bacterial

RNase H inthe presence ofpoly(dT) removed its poly(A) without significantly

affecting itshydrodynamic or electrophoretic properties in denaturing solvents.

Theviral 60 to70SRNAcomplexwas found to consist of 30 to40S subunitsboth

with (65%) and without (approximately 30%) poly(A). The heteropolymeric

sequences of these two species of 30 to 40S subunits have the same RNase

T,-resistant

oligonucleotide composition. Some,perhapsall, RNase

T,-resistant

oligonucleotidesof 30 to40S Roussarcomavirus RNAappeartohavea

unique

lo-cationrelativeto the

poly(A)

sequence, becausethe

complexity

of

poly(A)-tagged

fragments of30 to40S RNA decreasedwithdecreasingsize ofthe fragment.Two

RNase

T,-resistant

oligonucleotides which distinguish sarcoma virus Prague B

RNA from that of atransformation-defective deletion mutant ofthe samevirus

appear tobe associated with an 11Spoly(A)-tagged fragmentofPragueB RNA.

Thus RNA sequences concerned with cell transformation seem to be located

within 5 to 10% ofthe3'terminus ofPragueBRNA.

Poly(A) stretches have been found in the

RNA of tumor viruses and many cytocidal viruses as well as in many cellular messenger

and

heterogenous

nuclear RNAs(1, 3, 7, 13, 14,

16-18, 21-23, 26, 27,33, 35, 37, 45,47).Although the function of poly(A) in these RNAs is

un-known, it has been suggested that poly(A) is

added tomRNAs

post-transcriptionally (7, 34).

Consistent with this notion, the

poly(A)

stretches have been found at the 3' end of a

number ofmRNAs (31, 32, 40).

Based on endgroup-labeling techniques tumor virus RNA wasreported toterminate at

the 3' end with a

poly(A)

segment about 30

residues longin the case of an avian virus (43)

and with a

poly(A)

segment of 190 residues in

thecase ofmurinesarcoma-leukemiavirus(38). In this report we present evidence in agree-ment with earlier studies (23) that the poly(A) sequence ofRous sarcoma virus (RSV) isabout 180 nucleotideslongand confirm its location at or near the 3' end

by

two methods:

(i)

poly(A)

prepared enzymatically from RSV RNA

con-tainedoneadenosineper180AMPresidues. (ii) Removal of thepoly(A) stretch fromRSV RNA

by digestion with bacterial RNase H in the

presence ofpoly(dT) resulted in 30 to40S RNA

which was only marginally smaller than 30 to

40SRNA containingpoly(A). Further, no

inter-nal poly(A) sequences were found in 30 to40S RSV RNA.

In agreement with earlier observations on RSV RNA (23) and with a recent report on murine tumor virus RNA (20), we found that

about30%ofthe30 to40S RNAs ofRSV hadno

poly(A).Finger printanalysesof 30to40S RSV RNAs with and without poly(A) indicated that the two species are indistinguishable with re-gard to their heteropolymericsequences.

Moreover, fingerprint

analyses

of

poly(A)-tagged fragmentsof 30to40S RSV RNA showed

a decreasing

complexity

with

decreasing

size.

This indicates that locations of some, perhaps

all, RNase

T,-resistant

oligonucleotides

and of

polv(A) are the same on all 30to 40S RNAs. MATERIALS AND METHODS

Reagents.Thefollowingreagentswerepurchased.

[3H]uridine (40Ci/mmol) wasfrom NewEngland Nu-clear Corp.; [3H]adenosine (10 Ci/mmol or 16 Ci/ mmol was from New England Nuclear Corp. or Schwarz/Mann Research; Carrier-free 32p was from ICN; [3H]poly(A) (73

gCi/gumol

of phosphate) and poly(dT) were from Miles Laboratories; oligo(dT)-1515

on November 10, 2019 by guest

http://jvi.asm.org/

cellulose was from Collaborative ResearchInc.; RNase A, RNase

T,,

DNase I, and E. coli alkaline phospha-tase were from Worthington Biochemical Corp.; and RNase T2 was from Calbiochem. E. coli RNase H was prepared asdescribed previously (24).

Virus. Prague RSV ofsubgroup B (PR RSV-B) was used in all of these studies. It was propagated and purified according to published procedures (10, 11).

RNA preparations.Tobacco mosaic virus (TMV) RNA, [3H]uridine, [3Hladenosine,or32P-labeled 60to 70S RSV RNAs as well as30 to 40S subunitswere prepared according to published procedures (9, 11). All RSV RNAs used in these studies were isolated from radioactive virus harvested from infected cul-turesat3-h intervals.

[3HJuridineor32P-labeledchick cell 4, 18, and 28S RNAs were prepared as follows. Primary chick em-bryo fibroblastcells wereseededat aconcentrationof 4 x 101cells per 10-cmpetridish andweregrown in medium 199 supplemented with 2% tryptose

phos-phate broth (TPB), 1%calf serum, 1%chick serum, 1%dimethylsulfoxide;0.05%glucose,100unitsper ml ofpenicillin, 50Agperml ofstreptomycin, and 0.5 gg per ml offungizone. Twelve hourslater, mediumwas changedto8-ml per dish of Dulbecco modifiedEagle

medium with the same supplements as described aboveexcept that TPBwasomitted anddialyzed calf and chick sera were used. A 200-uCi amount of [3HJuridine or 1 mCi of 32P was added per dish; generally, fourdisheswerelabeledat onetime. After 10to12 h ofincubationat 42 C,disheswereremoved from the incubator and placed on an ice bath. Mediumwasremoved and the cellswerewashed twice with Tris saline; then, 1.5 ml of hypotonic buffer

containing0.01MTris-hydrochloride, pH7.4,0.01M NaCl, 2mM EDTA, and 0.05% Triton X-100 was addedtoeach dish. Aftersittingfor5minonice,cells wereharvested witharubberpolicemanandpipetted into aDouncehomogenizer.Cellswerebroken up with sixto sevenstrokes and under theseconditions,very few nuclei were broken. Nuclei were pelleted by

centrifugationat 600xg for5minin aSorvall centri-fuge. Thesupernatantwasthendiluted with standard buffer (0.01 M Tris-hydrochloride, pH 7.4, 0.1 M NaCl, and 1 mMEDTA), and the RNA extractedas described above for viral RNA. Afterprecipitationthe RNAwaspelleted and washed twice with 75% ethanol andredissolved in 0.3 ml of standard buffer contain-ing0.2%sodiumdodecylsulfate (SDS).Thesolution was heated at 100 C for 1 min, quicklychilled, and wasthen layered on a5-ml 10 to 25%linear sucrose gradient containing standard buffer plus 0.1% SDS. Sedimentation was in aSpincoSW65rotor at65,000 rpm for2 h at 20C. Fractions(0.3 ml) were collected and asmallportionofeachfractionwascountedin 3 ml oftoluene-based scintillation fluidcontaining 10% NCS (Nuclear Chicago). The sedimentation profile showed three distinct peaks representing 4, 18, and 28S RNAs.Peak fractions ofeach RNA were pooled, ethanol precipitated, and redissolved in buffer con-taining 0.01 M Tris-hydrochloride, pH 7.4, 10 mM NaCl, and1 mM EDTA. ThepurifiedRNAsamples were storedat -70C.

Conditions for enzyme reactions. DNase I:0.01M Tris-hydrochloride, pH 7.2, 4 mM MgCl2 and 20,ig

per ml of DNase, incubated at 38C for 30 min. Combined digestion with RNase A and RNase T,: 0.01MTris-hydrochloride, pH 7.2, 0.3 M NaCl, 1 mM

EDTA, 20ug perml of RNaseA, and150units perml ofRNaseT, incubatedat38C for 30 min. RNaseT,

alone: 0.01 M Tris-hydrochloride, pH 7.2, 0.15 M NaCl, 1 mMEDTA, and 150unitsperml ofenzyme incubatedat38C for60min. RNase T2 alone:0.04M ammonium acetate, pH 4.4, 1 mM EDTA, 5 units per ml of enzyme, incubated at 38C for2h. Combined digestion with RNase A,T,andT2: same conditions asforRNaseT2 alone plus20flgpermlof RNaseA, and 150 units per ml of RNaseT,,incubationat38C for2h. E. coli RNase H (24): reaction mixture (100 or 150uliters) contained 0.02 MTris-hydrochloride, pH 8.0, 10 mM MgCl2,6 mMdithiothreitol,25units per mlof enzyme and 350 pmol (- 1.85 x 104counts/min) of[3H]poly(A) plus175pmol ofpoly(dT). When RSV RNA (-0.4 Mg) was used as substrate, 0.25 gg of poly(dT) was added, whereas [3H]poly(A) poly(dT) wasomitted. The reaction was carried out at 38 C for 1h.Dephosphorylation and fractionation of commer-cial [3H]poly(A): 25 gtg (3.7 x 106 counts/min) of [3Hlpoly(A) (Miles) was treated with E. colialkaline phosphatase in a 1-ml solution containing 0.05 M Tris-hydrochloride, pH 8.2, and 2 units of enzyme. Afterincubationat38Cfor30min,itwasdiluted to4 ml with standard buffer containing 0.1% SDS and phenol-extracted as described for viral RNA. After ethanol precipitation, the poly(A) was pelleted and washed three times with 75% ethanol. It was redis-solved in 0.5 ml 0.01 M Tris-hydrochloride, pH 7.2. Forfurther fractionation aportionofpurifiedpoly(A) was heat denatured (100 C 1 min) in 0.3 ml of standard buffercontaining 0.2% SDS and was sedi-mentedthrough a5-ml10 to25%sucrosegradient as describedabove. Sedimentation wascarried out in a Spinco SW50.1 rotor at49,000 rpm for12 hat 20C. Peak fractions ofpoly(A)werepooled,ethanol precip-itated, and resedimentedsimilarly. Three subsequent sedimentations were performed to prepare poly(A) with a uniform sedimentation profile and a peak at about8S.

Isolation of[3H]adenosineor32P-labeled poly(A) from RSV RNA. Purified 60-70S RSV RNAin(<0.2 ml) 0.01 M Tris, pH 7.2, 1 mM EDTA was heat denaturedat100 Cfor 1minanddigestedeither with RNase A plus RNase T, or RNase

T,

alone as described above. Afterdigestion,thereactionmixture wasdiluted with standard buffer to 4 to5 ml after addition of 20Mg of carrier TMVRNA and SDS to 0.1% thesolution was phenol-extracted three times. The RNA was ethanol precipitated twice to remove all soluble nucleotides. Such poly(A) was either used directly forend-group analysisorfurtherpurified by heating (100 C, 1 min) followed by sucrose gradient sedimentationin anSW50.1rotor at49,000 rpmfor10 hat 20C.Peakfractionsrepresenting more than 80% of thetotalradioactivitywerepooled, andpoly(A)was then ethanol precipitated twice and used for end-group analysis.

End-group analysis. Purified commercial (Miles, Inc.) 7.8S, [3H]poly(A) (0.7to 0.9Mg) or [3H]adeno-sinelabeledpoly(A)from(0.5-1.5Mg)60 to70S RSV RNAwascompletelydigestedtomonoculeotideswith J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

RNase T2 in a volume of 0.2 ml (see above). The reaction mixture wasthen lyophilizedand redissolved in 30Mlitersof water. A200-ggamount of an equimo-lar mixture of each of the fourribonucleoside mono-phosphates and adenosine in 12 hliters buffer was added to it as internal standards. The mixture was subjected to electrophoresis on a Whatman 3 MM paper (12 by 56 cm) according to published proce-dures (23) except thatelectrophoresis was at 1,500V for 2.5 h. Afterthe paper had been air dried the spots representing adenosine and thefournucleotides were located by exposing the paper to UV light in a dark room. Each spot as well as regions between spots were cut outseparately. The nucleoside or nucleotides were elutedfromthepaper with0.4 mlofwater.Theeluted radioactivitywasmeasuredin15 mloftoluene-based scintillation fluid containing20%NCS andcounted in a Packard liquid scintillation counter. Afterelution, the paperwastested foranyremainingradioactivity

by the same method.Sincenoradioactivity could be detected on theeluted paper, the recoveryappeared

tobecomplete.

Oligo(dT)-cellulose column fractionation of

poly(A)-containing RNA. The procedures were a modification of the method of Aviv and Leder (4). One-half gram (dry weight)ofoligo(dT)-cellulose was washed and soaked in 0.01MTris-hydrochloride (pH

7.4) buffer for several hours at room temperature. Afterdecanting the fineparticles severaltimes,itwas packed intoacolumn (0.7by3cm). Before each run, the column was washed with 2mlof 0.1MKOH and subsequently washed with ice-cold buffer I (0.5 M LiCl, 0.01 MTris, pH 7.4, 0.05% SDS)until thepH wasneutral.AnRNAsamplein 0.2ml ofcoldbufferI wasapplied to the top of the column andallowedto enterthe column. Fractionation was carried out by

stepwise elution with 2-ml batches ofbufferI,buffer II containing0.1 M LiCl, but otherwise thesame as bufferI,8andbufferIIIcontaining0.01M

Tris-hydro-chloride and 0.05% SDS. Fractions (0.5 ml) were collected. The RNA ineach fractionwasdetermined

byitsradioactivity.

Milliporebindingand elution. Theprocedureisa slight modification of that described byMendeckiet al. (31).Briefly,theRNAwasdissolvedin/ordiluted atleast20-fold with buffercontaining0.5MKCI,0.01 MTris-hydrochloride (pH7.4), and1mMMgCl2and filtered slowly by gravity (approximately 6 ml/h)

through a Millipore filter previously soaked in the same buffer for 30 min. After washing three times with1-mlportionsofthesamebuffer,the RNAbound wasrecoveredby elutingwith three0.5-mlportionsof buffer containing0.1M Tris-hydrochloride (pH9.0) and0.5% SDS.

RESULTS

Sizeofpoly(A)inRSVRNA.Earlier studies

inthis (23) and other laboratories (16, 17) have estimated the size of the

poly(A)

stretch of tumorvirus RNA to be around 60,000daltons.

However, more

recently

the

poly(A)

stretch of

avianmyeloblastosisvirusRNAwasreportedto

be only 9,000 daltons (43). Since both the size

and the absolute amount of poly(A) in viral RNA must be known to distinguish between several relatively small or one relatively large poly(A) stretch, we have redetermined the size of thepoly(A) in RSV RNA.

Previous size estimates were done with poly(A) prepared from avian tumor virus RNA by RNase A and RNase

T,

digestion at high

ionic strength (see abovereferences). Although

poly(A) is known to be relatively resistant to RNase A athigh ionic strength (5), it is conceiv-able that the enzyme mayintroducesome nicks

inthis condition. This was indicated by

experi-ments in which poly(A) prepared from viral

RNA by digestion with RNase T1, which is

specificforG,andwithacombinationofRNase

T,

and RNase A were compared. It was found

that after electrophoresis in formamide-polyac-rylamide gels the RNase T1-prepared poly(A)

wasmuchmorehomogenous than thatprepared

earlier byRNase A (Fig. 1, ref. 23). If

commer-cialpoly(A) wastreated under the same

condi-tions as used to prepare viral poly(A) with RNase T1, its electrophoretic distribution was the same as that of an untreated control.

However, commercial poly(A) was rendered

heterogenous if treated with RNase A and

RNase

T,

(notshown). We conclude that RNase

A at high ionic strength nicks viral and other

poly(A) sequences. Nicking during isolation by

RNase A may be the reason forthe size

differ-ence between the

poly(A) prepared by

us and

that described

by Stephenson

et al. (43) and

may also account for the heterogeneity of

A-rich fragments rangingfrom

poly(A)

tooligo(A)

describedby Horstetal.(19).

The molecular

weights

of commercial 7.8S

poly(A) and of the poly(A) of RSV RNA

pre-pared by RNase

T,

were estimated from their

electrophoretic mobilitiesin

formamide-polyac-rylamide gels to be 116,000and70,000,

respec-tively (Fig.1 insert), using18S rRNA (0.7 x 106

[42]) and tRNA (2.5 x 104 [42]) as standards

and assuming a linear inverse relationship

be-tweenthe

logarithm

ofthemolecularweightof a

polynucleotide

and its

electrophoretic

mobility

(42).However, this sizeestimateof

poly(A)

may

be subject to 10 to 20% error since deviations

from a linear log molecular weight-mobility

relationship have been observed under similar

conditions (12).

Further it is conceivable that viral poly(A)

prepared by RNase

T,

contains at its ends a

stretch of nucleotides containing U, A, and C which would notbe digested byRNase

T,.

This may affect slightly the molecular weightofthe

poly(A)

prepared by

RNase

T,.

However,

base analyses of viral poly(A) prepared with

on November 10, 2019 by guest

http://jvi.asm.org/

I

'-4

x

0s

0

_=L

3

-0

x

0

0-CM

#LzL4~ <4L2:

0

20

40

60

80

100

[image:4.499.118.402.61.368.2]

DISTANCE MOVED

(mm)

FIG. 1. Determinationofthe sizeof marker 7.8Spoly(A)andRSVpoly(A) by formamidepolyacrylamide gel electrophoresis. RSVpoly(A)wasisolatedfrom30 to40S [32P]RNA subunitsby RNase

T,

digestion. Thedigest wasphenol-extractedthreetimes and ethanol-precipitatedin thepresence of 30

,g

of TMVRNA. The RNA recovered was then passed through anoligo(dT) -cellulose column (Fig. 3). The poly(A) peak eluted with buffer IIIand wasdivided into twoportions. One portion was ethanol-precipitated for analysis of its base composition asdescribed for Table 1. The other portion was mixed with 3H-labeled 4S, 18S RNAs, and 7.8S poly(A). After ethanolprecipitation the mixturewasdissolved in buffered formamide and subjected to electrophoresis in a 5% polyacrylamidegelcontaining98%formamideat 100Vfor 7 hasdescribedpreviously (12). The insert shows a log molecularweight/electrophoreticmobility plot (42) used to estimate the molecular weight of RSVpoly(A).

RNase

T,

and a combination of RNase A

and RNase

T,

werevery similar andsuggested

that very few (<3%) bases other than A were

present inour

poly(A)

(not shown). Weassume

that some ofthose bases come from the5 end

of RNase

T,-prepared

viral poly(A). Thus it

appears that our size estimate of viral poly(A)

was little, at most 3%, affected by bases other

than A. Moreover, we may conclude that, in contrast to anearlier report(19), the poly(A) of RSV RNA is free of interspersed G-residues sensitive to digestion by RNase

T,,

since it is

notdegraded bythis enzyme.

Locationof thepoly(A)at or near the 3' end

ofRSVRNA.There issuggestive evidence from end group analyses that the poly(A) of tumor virus RNA, like the poly(A) of other viral or

cellular mRNAs is located at the 3' end of the

RNA molecule (2, 38, 43, and J. Keith, M.

Gleason, and H. Fraenkel-Conrat, Proc. Nat.

Acad. Sci. U.S.A., in press). One method used

todetermine the3' terminallocation ofpoly(A)

was based on oxidation and reduction of the

vicinal 2' and3' OH groups at the 3' terminus

ofthe RNA (38, 43). However this method has

ledto contradictory resultswhen appliedto

tu-morvirusRNA (15, 25, 30).

Another method was based on the ratio of

nucleosides to nucleotides after complete

hy-drolysis of radioactive RNA or poly(A) (2, 38,

and J. Keith, M. Gleason, and H.

Fraenkel-Conrat, Proc. Nat. Acad..Sci. U.S.A., inpress).

However, the accuracy of this method suffers

from the low

nucleoside/nucleotide

ratios of

RSV RNA (1:200 for poly(A), 1:10,000 for 30

to 40S RNA). We describe below two

ap-proaches to locate the poly(A) stretch in RSV

RNA.

End-group analysis ofviral poly(A). First

we determined the end-group of biologically

labeled poly(A) from 30 to 40S RSV RNA. If

poly(A) is at the 3' terminus of the RNA it

1518

on November 10, 2019 by guest

http://jvi.asm.org/

should, after complete digestionto mononucleo-tides, containoneadenosineperabout 200AMP residues, assuming our size estimate was

cor-rect. We have chosen enzymatic digestion of

poly(A) instead ofdigestion by alkali, because

pilot experiments had shown that approxi-mately 50% ofthe 3H of poly(A) was rendered

volatile during lyophylization after KOH hy-drolysis, presumable due to isotope exchange with water (39, 46). Exhaustive digestion with RNase A also proved unsatisfactory, as the

digestion productwasmostly ApAprather than

AMP (not shown). Complete digestion of viral,

as well as commercial poly(A) to mononucleo-tideswas obtained with RNase T2 (Table 1).

It is shown in Table 1 that the poly(A) from

RSV RNA contained one adenosine per 180

AMP residues and that commercial poly(A) containedone per320. This is compatible with our size estimate of about 200 nucleotides for

viral poly(A) and with the size estimate of 116,000for commercial 7.8S poly(A).

Nosignificant radioactivitywasdetectableat the origin of the pherogram usedto resolve the digest (Table 1), neitherwasthereradioactivity

between the spots of the four UV-absorbing monophosphate ribonucleotide markers usedto monitor the electrophoresis. However, there

weresmallamountsofradioactivity in the CMP spot if viral or commercial poly(A) were

ana-lyzed (experiments 1 and2, Table 1). Thiswas

most likely due to slight trailing of AMP,

because afterprolongedelectrophoresis

(experi-ment 3, Table 1) no radioactivity was seen in

the CMP position. Significant amounts of

ra-dioactivity migrated with the GMP spot, when

the poly(A) ofRSV was analyzed, especially in

experiment 3 (Table 1). This could be due to

residual heteropolymeric sequences associated

with this preparation of viral poly(A) or less

likely (see above; 38) due to RNase

TI-resis-tant G residues interspersed into poly(A). The

radioactivity inGMPmayhave originated from

cellular conversion of [3H ladenosine into

[3H ]GMPorfrom [3H ]guanosine contaminating

the [3H]adenosine used by us (from New

Eng-land Nuclear Corp., experiments 1 and 2,

Ta-ble 1; from Schwarz-Mann, experiment 3, Table 1). The amount of 3H electrophoresing with GMPvaried relativetothat found inAMP

in the experiments described in Table 1. By contrast, the radioactivity associated with the

adenosine spot.relative to that in the AMP

spot remained constant (.55%) in all experi-mentswithpoly(A) fromRSV(Table 1).

Control experiments using 28S [32P]rRNA (1.25 x 105 counts/min) showed that <0.04%

(<50 counts/min) of the mononucleotides had

beendephosphorylated underour conditionsof

enzymatic digestion using RNase T2, RNase

TI,

and RNaseA (Fig. 2). Thismaybeaproblemto be considered if alkali is used to digest RNA (28). Therefore, we believe that the

adeno-sine residues detected by complete enzymatic digestion of viral poly(A) werederived from the

3' end of the RNA, and consequently that

TABLE1. End-group analysis of marker and RSV poly(A)

Marker[3H]poly(A)" [3H]poly(A)from 60to70SRSV RNAb

Poly(A) tested I II III I II III

experiments

Counts/ % Counts/

%,

Counts/! Counts/! Counts/! Counts/

min' min min min min min

Adenosine 400 0.31 328 0.32 400 0.31 314 0.59 144 0.55 430 0.52

CMP 246 0.18 145 0.14 0 0 162 0.30 46 0.19 0 0

AMP 128,780 99.43 103,410 99.44 129,900 99.66 52,890 98.65 24.190 98.58 80,000 97.29

GMP 72 0.06 70 0.07 36 0.03 218 0.41 144 0.55 1,800 2.19

UMP 32 0.04 36 0.04 0 0 31 0.06 13 0.05 0 0

Total 129,530 100 103,990 100 130,340 100 53,165 100 24,540 82,230 100

AMP/adenosine 322:1 315:1 325:1 170:1 168:1 186:1

Avg ofAMP/ 320±5 178+8

adenosine

aComplete digestion of poly(A) by RNase T2 was described in Materials and Methods. In experiments I and II, dephosphorylatedandethanol-precipitated Miles poly(A)was used: inexperimentIIIpolv(A) hadbeen furtherpurified by

threesubsequentsedimentations. l

[3H]adenosine-labeledpoly(A) fromRSV RNAwasisolatedby digestionwith RNasesT,and Aasdescribed in Materials and Methods. In experiments I and II, poly(A) isolated from same RNA preparation was used directly without further

purificationbysedimentation. InexperimentIII, poly(A) isolated fromadifferent RNApreparationwasfurtherpurified by

sedimentation to remove the oligonucleotides that might have co-precipitated with poly(A). Paper electrophoresis of

experimentIIIwasrunfor 3h instead of2.5h.

cThirtycountsper minute ofbackground eluted frominterspot regionhas beensubtracted.

1519

on November 10, 2019 by guest

http://jvi.asm.org/

O

L

20 23

DISTANCE MOVED (Cm)

FIG. 2. Electrophoresis of the 32P-mononucleotides produced by complete enzymatic digestion of 28S chick rRNA. Approximately I 1sgof [32PV28S rRNA (1.2 x 105 counts/mm) was digested with RNases A,

T,,and T2 in a total volume of 0.2 ml at 38 C for 2 h. The digest was analyzed by paper electrophoresis. A

20-Muliter amount of [3H]32P04 (2 x 104counts/mmn)

was spotted on the same paper and run in parallel as a p043- marker. The32p nucleotides and pQ43 marker were detected by determining the radioactivity of 1-cm stripsof the pherogram in toluene-based scintil-lation fluid.

poly(A) appears to be located at the 3' end of viral RNA. However, our experiment does not rule out the possibility te the penultimate base(s) of viral poly(A) are not A but a few U, C, G residues resistant todigestion with RNasesTA

and A used to prepare our poly(A). Further, it may be argued that our evidence for the 3'

location of poly(A) is not compelling, since it is

based entirely on the end-group of the viral

poly(A) which represents only 0.5% ofthe

radio-activity associated with the poly(A) molecule. 30 to 40S RSVbefore and after removal of

poly(A)

with RNase H of E. coli. Therefore,

independent

evidence was sought to locate the

position of

poly(A)

on viral 30 to 40S RNA.

Using RNase H in the presence ofpoly(dT) to

digest specifically the

poly(A),

but not the

heteropolymeric sequences of viral RNA, it should be possible to

determine

thelocation of poly(A) on the RNA by analysis of the RNase H-resistant

heteropolymeric

sequences of the

RNA.

An RNase H which is free of other

nucleolytic

activities

is essential for this

experi-ment. Starting material for

these

experiments was RSV RNA,

[3Hsadenosine

or 32P-labeled, selectedfor an intact

poly(A)

stretch by binding and subsequent elution from

oligo(dT)-cellulose

at low

ionic

strength (fraction III, Fig. 3). This RNA was incubatedwith E. coli RNase H in the presence of

poly(dT) (24)

(see Materials and

Methods). The RNase H-treated RNAwasthen

compared to untreated RNA with regard to its

Millipore filter binding capacity, its poly(A)

content andsize.

Analysis of 30 to 40S RSV RNA before and

after digestion with RNase H and in the

pres-ence of poly(dT) indicated that 80% of the

RNaseA-and RNase

T,-resistant,

viralpoly(A)

can be digested by RNase H (Table 2). It

appears then thatmost orall poly(A) (about15

to 20 pmol) ofabout 1,200 pmol ofviral RNA

(specific activity -105

counts/min/,ug)

was

di-gested in our conditions. Control experiments

using the same amount of enzyme, indicated

that 110 to 140 pmol of

[3Hlpoly(A)

.poly(dT)

were digested in the same conditions. Thus,

digestion of the viral poly(A) was at enzyme

excess (Table 2).

Moreover

Millipore

binding capacity

of 30to

40S

RSV RNA was reduced from about 90 to

10%bytreatmentwith RNaseH(Table 2). The

enzyme/substrate

ratios in these experiments

were about the same as those used above

Isee

[image:6.499.64.253.60.239.2]

control experiments with [3H]poly(A).poly(dT),

Table 21.

The size ofthe RNase H-treated 30 to

40S

RNA was

indistinguishable

from that of

un-treated RNA on the basis ofsedimentation in

formaldehyde-sucrosegradients (Fig. 4) butwas

found to be slightly smaller on the basis of its

0

x 0 0L

cI

cL

30

20

10

0

I

II

'III

.5M

.1M Tris

K

1k1

I

L

0

4

8

12

16

FRACTION NUMBER

FIG. 3. Oligo(dT)-cellulose chromatography of 30-40SRSVRNA.Purified 30 to40S [32P]RNAwas dissolvedin 0.2 mlof bufferIandheatedat100 C for1 min. Fractionationwas asdescribedinMaterials and Methods. Arrows indicate where the elution with different buffers (I, II,andIII) started. Fractions (0.5 ml) were collected and

5-gsliter

portions of each fractionwerecountedin3mloftoluene-based scintil-lationfluid containing10%NCS.

1520

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.499.267.453.389.587.2]

[image:7.499.54.452.86.387.2]

TABLE 2. 30 to40S RSVRNAaandpoly(A) before and after digestion with RNase H Fraction resistant to RNases A +T,after:' Total 30 to 40S RNA

poly(dT) RNaseH RNaseH+poly(dT)

Expt Counts/min Label Counts/minc % Counts/min % Counts/min %

I 3710 [3H]A 235 6.1 44 1.2

II 4200 [3H]A 240 5.7 220 5.2 40 0.9

III 4100 [3H]A 260 6.1 245 6.0 75 1.7

IV 21,000 32p 345 1.6 95 0.5

Total [3H]poly(A) Fraction digested by RNase H with poly(dT)

Expt Counts/mmn pmol Counts/min pmol %

V 18,700 353 7,200 136 38

VI 18,500 349 5,800 109 31

Milliporebindingcapacity'

Expt Substance -RNaseH +RNase H

Counts/min Counts/min Counts/min pmol Counts/min pmol %

applied bound applied p bound

VII [32P130 to40S 3,788 3,297 87.3 5,200 574 11.0

RSV RNA

VilI [3HIpoly(A)e 13,800 260 5500 122 47.0

aPoly(A)-containing 30to40SRNApooledfrom fractionno.9 to11showninFig.3.

RNase H digestionwascarriedout asdescribed in Materials and Methodsusing approximately 0.4ug of

[3H]adenosine (1.5 x 104counts/min) or 32P (7 x 104 counts/min) labeled RSV RNAineach assay. Control tubes eitherminusRNaseH orminuspoly(dT) (0.25,g) but with RNase H were incubated in parallel. At the end of reaction, duplicate portions of reaction mixture from each tube were taken separately. They were incubated with DNasetoremovepoly(dT)afteradjustingtheir volumeto100

pliters

and the saltconditionsto those describedinMaterials and Methods for digestion with DNase. Subsequently they were adjusted to 0.5 ml pertube andsubjectedtoRNaseAplus RNaseT1digestionforthe assay of the remaining poly(A) sequences. At the endofdigestionby RNases, 100,gofyeasttRNAwasadded andtrichloroacetic acid precipitable material wasdetermined (44). Each number represents the averageofduplicate assays. Fifteen counts per minute of background had been subtracted.

cTheamountofRNA resistant toRNases.

dRNase Hdigestionwascarriedout asdescribed in ". A control without RNase H was incubatedinparallel. Each tube contained 1.5 x 104counts/minof32P-labeledpoly(A)-containing 30to40S RNA. After incubation eachreaction mixturewasadjustedto0.1 MNaCl and1 mMEDTA anda 100-fold excess ofcold poly(A) (25 Mg)wasaddedtopreventpoly(dT) from complexingtoRSV RNA. They were thenheat-denatured at 100 C for 40 sand chilled. Each reaction mixturewasadjustedto atotal volumeof 0.3ml withSDS standard buffer and sedimented through a5 ml 10 to25%linearSDS sucrosegradient afteradding 18S [3H]rRNA as an internal marker(Materials andMethods). Aftersedimentation, 30to40SRSV RNAswerepooledseparately, ethanol-precipitated andsubjectedtoMilliporebinding(MaterialsandMethods).

eStandard assaysofRNaseHwith [3H]poly(A) andpoly(dT) wereperformedinparallel and carried outas described in Materials and Methods. Eachnumber represents theaverage of duplicate assays.

electrophoretic mobility in

formamide-polya-crylamide gels (Fig. 5). This suggests that the

poly(A) segment is locatedatornearaterminus

ofthe RNA, since its removal did not

signifi-cantly affect the size of the RNA. Given a

terminal location of the

poly(A),

we would expect at most a small increase in electropho-retic mobility after removal of

poly(A),

since

thepoly(A) segment represents onlyabout 1.5% ofthe 30to40S RSV RNA.Further,therather

homogenous distribution ofRNaseH-treated30 to40S RSV RNA informaldehyde gradients or formamide gels indicates that the RNase H usedby us wasfree ofdetectableendonuclease.

The second minor peakofRNA observedboth in the RNase H-treated and untreated RNA

1521

on November 10, 2019 by guest

http://jvi.asm.org/

I 28S - NsH 28S + 2S

RNaseH I +RNase H Complete

cl 4 o poly(dT) poly(dT) ( an

RNase H(C)asdsrbdi al .Sbeuety ahsml a iue it tnadbfe o2m fe

X 3 -x

addingtoeah1C-)ro f02MDAan 0y fTVN.Atrpeoletato 8 SPr a

21-0 I 0

016e04 8 12 4 8 12 160 4 8 12 16

FRACTION NUMBER

FIG. 4. Formaldehydesucrosegradientsedimentationof30 to40SRS V RNAbeforeandaftertreatmentwith RNaseH.Poly(A)-containing [3H]adenosine-labeled30 to40SRSVRNAwaspreparedasdescribedfor Fig. 3. Three identicalportions of30 to 40S RNA were incubated with poly(dT) (A), RNase H(B), poly(dT), and RNase H(C)asdescribed in Table2.Subsequently, eachsamplewasdiluted with standardbufferto2mlafter addingto each 10 litershof 0.2MEDTAand 30 ugofTMVRNA.After phenolextraction28S [3eP]rRNAWas

added toeachsample; the RNAs werethen ethanolprecipitated. After pelletingandwashingtwice with 75% ethanol,the RNA ineachtubewasredissolved in0.3mlofbuffer containing0.1MNaCl,2mMEDTA, pH 7.2, 0.1%SDS,and1.1Mformaldehyde. Afterincubationat65 Cfor15mineach RNAsamplewaslayeredonthe top ofa 5-mI10 to20%sucrosegradient containingthesamesalts andformaldehydeas the bufferdescribed above. SedimentationwasinparallelinaSpincoSW65rotor at65,000rpmfor3hat20 C.Fractions(0.3 ml)

were collected through the bottom of the tube, and

100-Mgliter

portions of each fraction were measuredfor radioactivity.

control seen in the

formamide-polyacrylamide

gels (Fig. 5) is

thought

to be 30 to 40S RNAof

size class

b,

derivedfrom

transformation-defec-tive virus which segregates

spontaneously

from stocks ofnondefective aviansarcomavirus(29).

We conclude thatRNase H treatment

effec-tivelyremovesallor mostofthe

poly(A)

associ-ated with viral RNA without

significantly

re-ducing its size. It follows that the

poly(A)

is

located near a terminus and that there are no

largepoly(A) stretches located within the30to

40SRNA chain.

Notall30 to40S RNA subunits of60to70S

RSV RNA contain

poly(A). During

this and

earlier (23) studies, it was repeatedly observed

thatheat-denatured 60 to 70S RNAorpurified

30 to

40S

RNA had a much lower

binding

capacitytoeither

Millipore

filtersor

oligo(dT)-cellulosethan intact60 to70S RNA. This raised

the question whether all subunits in 60to 70S RNA contain a poly(A) sequence. It seemed

plausible that if there exist some 30 to 40S

subunits which do not contain poly(A) se-quences, they could bind tooligo(dT)-cellulose

only ifcomplexed toother, poly(A)-containing

subunits witnin the same 60 to 70S RNA

complex. Therefore, the

oligo(dT)-cellulose

binding capacityofpurified60 to70S and 30 to

40S RNAs were compared at different ionic

strengths. The results, summarized inTable3,

indicate that 90% of the60 to70S butonly65%

ofthe30 to40S RNA bindtooligo(dT)-cellulose athighionicstrength, implying that about1 out of3 30 to40S RNAspecies isdevoidofapoly(A)

stretch that confers binding capacity for

oli-go(dT)-cellulose. This is in agreement with a

recent report by Ihle et al. (20) that only

two-thirds ofpurified 30 to40SRNA of mouse

leukemia virus RNA binds to

poly(U)-sepharose. As expected, most of an 18S

ribo-somal RNA used as internal control in the

binding

studies wasincapableofbindingto an

oligo(dT)-cellulose (Table3).

Direct analyses of the poly(A) content of

fractions of 60 to 70S and 30 to 40S RNA,

distinguished by

their differential binding

ca-pacities to oligo(dT)-cellulose, confirmed the

above conclusions. It was found that after

complete enzymatic

digestion

of

heteropolym-eric sequences with RNase A and RNase

T,

(Materials and

Methods),

30 to 40S and 60to

70S RNA eluting from oligo(dT)-cellulose at

high ionicstrength (fraction I, Fig. 3, Table 3) were devoid of enzyme-resistant RNA and therefore did not contain poly(A) (not shown). However, 60 to 70S and 30 to 40S RNA eluting atlowionicstrength (fraction

III,

Fig. 3, Table 3)contained1.5%resistantRNA,iflabeled with 32P, and 6% if labeled with [3H]adenosine as

expected for a poly(A) containingRNA (Table

2).The RNAspecieswhichelutedat intermedi-ateionic strength contained 1to 1.5%resistant

1522

on November 10, 2019 by guest

http://jvi.asm.org/

[image:8.499.113.396.81.222.2]

0 20 40 60

DISTANCE MOVED

(mm)

FIG. 5. Formamide polyacrylamide gel electrophoresis of 30 to 40S RSV RNA before (A) and after (B)

treatmentwith RNase H.Poly(A)-containing 32P-labeled30to 40S RSV RNAselutingat low ionicstrength from oligo(dT)-cellulose(fractionIIIshown inFig.3)wasincubated withRNase HasdescribedforFig.4.After phenolextraction, [3H]adenosine-labeledpoly(A) containing 30 to 40S RNA and 18SrRNA wereaddedas

internalmarkers. The RNAswereethanolprecipitated.Electrophoresiswasinpolyacrylamidegelscontaining

98%formamideat 100 Vfor 12hasdescribed (12).

RNA, if labeled with [3H ]adenosine (not shown). The degree ofRNase resistance in this RNAspeciesvaried with different preparations,

sothisfraction of RNA islikelytoinclude RNA

species carrying poly(A) segments which are

much shorter than those of the RNA fraction

eluting from oligo(dT)-cellulose only at low ionicstrength (fractionIII, Fig. 3).Weconclude that 60 to 70S viral RNA contains 30 to 40S

RNAsubunits withpoly(A) (-65%), some

with-out poly(A) ("30O%) and a few subunits (5 to 10%) with short segments ofpoly(A).

The sizes of 30 to 40S viral RNA subunits which lack poly(A) (fractionI, Table3, Fig. 5)

were compared to the 30to 40SRNAsubunits which contain poly(A) (fraction III, Table 3, Fig. 3) by simultaneous electrophoresis in form-amide-polyacrylamide gels. It can be seen in

Fig. 6A and B that 30to40S [3H]-or [32P]RNA

withpoly(A) migrated alittle slowerthan 30to

CV

0

V'-4 x

Q-(O4

8

4

0

x

0 CL 0

Q r)

15'):3

on November 10, 2019 by guest

http://jvi.asm.org/

[image:9.499.102.388.64.482.2]

[image:10.499.57.454.74.205.2]

TABLE 3. Oligo(dT)-cellulose column binding of RSV RNAa

60to70S RNA 30 to40S RNA 18SrRNA5 Experimentsand RNA Labeling %Eluted %Eluted %Eluted

preparations

IP [3HlU 5.7 4.3 90.0 18.0 24.0 58.0 86.0 5.6 8.4

32p 5.0 4.0 91.0 14.0 18.0 68.0 87.0 5.8 7.2

II [3HJA 5.4 3.6 91.0 17.6 15.7 67.0

III [3H

]A

29.5 8.5 62.0

IV 32P 25.0 13.0 62.0

Vd 32P 21.8 5.2 73.0

Average 5.4 4.0 90.7 21.0 14.1 j 65.3 86.5 5.7 7.8

Purified RSV60 to70S or 30 to 40SRNAsweresubjectedtooligo(dT)-cellulosecolumn fractionationas de-scribedin Materials and Methods andforFig.3.Anelutionprofileof 30to40S is shown inFig. 3.Percentages ofradioactivityelutingathigh,intermediate, and low ionicstrengthweretermed I,II,and III,respectively, and correspondtotheelutionprofileshown inFig. 3.

h [3H]uridineor32P-labeled 18SrRNAswereusedasinternalcontrols in the elution of 60to70S RSV RNAin Exp. 1.

[3H]uridineand 32P were addedtodifferent dishesofthe samevirus-infected culture and the RNAswere isolated inparallel.Thesedif'ferentially labeled RNAswereusedforcomparisonoftheir size(Fig.4and 5).

dFractionated30 to40S RNAsofthisexperimentwereusedforfingerprintanalysis(Fig. 7).

40S [3H]-or [32P]RNA withoutpoly(A). An18S

rRNAmarker isincluded in these experiments;

in addition, a shoulder is seen at the leading edgeof eachpeak,thisisthoughttobe30to40S RNA of size class b (see above). A control

experiment showsthatpoly(A) containing30to

40S [3H]- and [32P]RNA coincide upon

electro-phoresis in the same conditions (Fig. 6C). It

appearsthen,that in agreement withthe exper-iments described above in which poly(A) was

removed enzymatically from 30 to 40S RNA,

the presence of poly(A) reduces slightly the

electrophoretic mobilityof 30 to 40SRNA.

Theexistence of 30to40SRNAsubunitswith and without poly(A) in 60 to 70S tumor virus

RNA observed here and by others (20, 23),

raises the question ofwhether

poly(A)-contain-ing andnonpoly(A)-containingRNAsrepresent

genetically different fractions of the RNA. To

test this the fingerprint patterns of RNase

T,-resistant

oligonucleotidesofthetwotypes of

30to40SRNAwerecompared. Itcanbeseen in

Fig. 7 that, except for the absence of the

poly(A) spot (lower right corner in Fig. 7A) in

thepoly(A) lacking30 to40S RNA, the

finger-prints ofpoly(A)-containing and poly(A)-lack-ing30to40SRNAspeciesareidentical. There-fore, we conclude that the 30 to 40S RNA species with and without poly(A) stretch are

genetically identical. This conclusion is

com-patible with thefinding that the complexity of

aviantumorvirus RNA is onthe order of 3.5 x

106daltons, whichistheapproximate

complex-ity ofone 30to 40Ssubunit only (K. Beemon,

P. Duesberg, and P. Vogt, Proc. Nat. Acad.

Sci.U.S.A., inpress).

Locations of some RNase

T,-resistant

oli-gonucleotides and of poly(A) appear to be the same in all 30 to 40S RNAs. The above

experiments and those of others (43) indicate

thatpoly(A)is locatedat or nearthe 3' end of 30

to 40S RSV RNA. If the heteropolymeric se-quences of 30 to 40S RNA all had the same location on the

polynucleotide

relative to

poly(A), instead ofbeing circularly

permuted,

poly(A)-tagged natural or artifically

produced

fragments would be very useful for sequence

analysesofviral RNAs. It would beexpected,for

example, that ifthe

heteropolymeric

sequences

of all 30 to 40S RNAs are located

identically

relativetopoly(A), the

complexity

ofa

poly(A)-tagged fragment of 30 to 40S RNA should be

directly

proportional

to the size of that

frag-ment. However, if the RNA sequences are

extensively

permuted

the

complexity

of a

poly(A)-tagged

fragment must notbe less than

that of the total RNA. To

distinguish

between thesealternativeswehavealkali-degraded 30 to

40S

[32P]RNA

into fragments sedimenting

broadly with apeakof11S (36), corresponding

to anaverage size ofabout240,000daltons(41).

The poly(A) containing fragments of the

de-graded RNA were isolated by binding and

subsequentelution fromMilliporefilters

(Mate-rials and Methods). The recoveries ranged

be-tween 5 to 10%indifferent experiments.This is

expectedifone considers that an 11S fragment

of240,000 daltons corresponds

roughly

to 5 to

DUESBERG 1524)

on November 10, 2019 by guest

http://jvi.asm.org/

~~~~~~~1-.-v,-\,.. .. .,

15 18 (^) Fraction Il-1

04~ ~~ab

o0

4 20

phrei. 30t 0 N ihutpl()wspe

0 ~~~~~~~~~~~~~~~0~

C 30-40S Peaka

15

-l(a)FractionIII 15

18S

pl( Fraction III

10 b 10

5 -5

0l

0

0 20 40 60

[image:11.499.51.237.71.547.2]

DISTANCE MOVED(mm)

FIG. 6. Comparison of the size of poly(A)- and

non-poly(A)-containing30 to 40S RSV RNA isolated

from 60 to 70S viral RNA byformamide gel electro-phoresis. 30 to 40S RNA without poly(A) was pre-pared byelutionfromoligo(dT)-celluloseathighionic strength (fraction I, Fig.3) and30to40Swithpoly(A)

was thatwhich elutedatlow ionic strength (fraction III,Fig. 3).Portionsof30to40SRNAfrom

oligo(dT)-cellulosefractions IandIIIweremixedtogetherwith

18S3rRNAandanalyzed by formamide gel

electropho-resisas describedforFig.5. (A) 30to40S RSVRNA withoutpoly(A) (fraction 1)and 30to40S RNA with

10%ofa 30 to

40S

RNAof 3.5 x 101daltons.

We have compared the complexity of these

poly(A)-tagged fragments of 30 to40SRNA to

that of undegraded 30 to 40S RNA by

finger-printing. It can be seen in Fig. 7C that the

numberofRNaseT1-resistant oligonucleotides,

and thus the complexity ofpoly(A) containing

11S fragments of PR-B RNA, is significantly

lower than that of 30 to 40S RNA (Fig. 7A).

Many large oligonucleotidesseen in the

finger-print of 30 to 40S RNA are completely absent

from the pattern of poly(A)-tagged 11S RNA

fragments. These arethoughttobe

oligonucleo-tides located most distantly from the poly(A)

stretch on the 30 to

40S

RNA. Some

oligonu-cleotides marked

by

letters in

Fig.

7AandCare

present athigh and some atintermediate con-centrations relative to the total counts per

minute on the fingerprint or to the counts per

minute ofthepoly(A) spot(Table4).Theseare

thought to be located at or near the poly(A)

stretch of 30 to 40S RNA. Varying relative

concentrations ofoligonucleotides are expected from the heterogenous size distribution of the

poly(A)-tagged fragments used to prepare the

fingerprint pattern. Since fragments of larger

size would contributeoligonucleotidesnot

pres-ent in smaller fragments. We conclude that

some, perhaps all, RNase

T,-resistant

se-quences of tumor virus RNA have the same

location relative tothe poly(A).

Itshould then be

possible

tomapthe position

of large oligonucleotides relative to that of

poly(A) on the 30 to 40S

polynucleotide

by

comparing the mass ratios ofa given

oligonu-cleotide tothe total RNAor tothatofpoly(A)in

the fingerprint patterns derived from intact

poly(A)-tagged 30 to 40S RNA and

poly(A)-tagged fragmentsofRNA. Forexample, it can

bededucedfrom the datainTable4thatspota

and b must be more distant from the poly(A)

stretchthan spot c.

DISCUSSION

The experiments described here confirm and

extend earlier studieson

poly(A)

in tumorvirus

RNA (16, 17, 23) andsuggestthattwo-thirdsof

the 30 to 40S subunits of 60 to70SRSV RNA contain a poly(A) stretch consisting of about

200nucleotidesat or nearthe 3' endofthe RNA.

The finding that some 30 to 40S RNA

sub-poly(A) (fraction III). (B) Different preparations of thesame twospecies of30to40SRSVRNAasin(A).

(C) Two preparations of30 to 40S RSVRNA with poly(A). Peakarepresents 30to40S RSV RNA. Peak b is the18S[3H]rRNAaddedas aninternal standard.

on November 10, 2019 by guest

http://jvi.asm.org/

-.

--0

-i be *

A

*

o

1~ f it

_

i~~

A. %

a,

ID0

[image:12.499.57.452.77.298.2]

B.0

C

l_

D

FIG. 7. (A-O) Fingerprint analyses of30to40SRSVRNA containingpoly(A), (A); of30to40SRSVRNA withoutpoly(A), (B);andofpoly(A)-tagged 11Sfragments of 30to40SRSVRNA, (C); and of the 30to40SRNA ofatransformation-defectivePRRSV-B(D).30to40S RSVRNA without poly(A)wasprepared by elution from

oligo(d7)-celluloseathighionicstrengthand 30to40SRSV RNA withpoly(A) byelutionatlowionic strength (see Fig. 3). The RNAs were exhaustively digested with RNase T1 and the digests were analyzed by two-dimensionalfingerprint analysisas described (K.Beemon, P. Duesberg, and P. Vogt, Proc.Nat. Acad.Sci. LS.A.,inpress).AnI S,poly(A)-containing fragment of30to40SRNAwasprepared byincubating 30to40S RSVRNA in50 Mliters 0.05 MNa2CO,(pH 10)at50 Cfor 6min (36). The solutionwas thenmixed with300

Aliters ofstandardbuffercontaining 0.2%oSDS andsedimentedinthepresenceofan18S

[9H]rRNA

standard.

Asymmetrical peak sedimentingat11Swasobtainedandethanolprecipitated. The pelletwasredissolvedand

passed throughaMilliporefilter.Millipore-bound 11SRNA(5 to10%of the 11S RNAapplied) waselutedas

described inMaterialsandMethods andfingerprintedasdescribedfor (A).

TABLE 4. Radioactivitya of certainoligonucleotidesand of the total RNA from which itwasderivedafter fingerprintingasshowninFig. 7aandc

60 to70SRNA Poly(A) fragment Poly(A) fragment

(Fig. 7a) (1OS) (11S) (Fig.7c)

Partofdigestanalyzed

Counts/min Counts/min % Counts/min %

Total 1.17x 106 100 48,750 100 70,400 100

Poly(A) 15,560 1.33 8,140 16.7 9,190 13.1

Spota 1,470 0.13 31 0.06 79 0.11

Spot b 2,900 0.25 48 0.10 108 0.15

Spotc 1,730 0.15 391 0.8 570 0.81

Spot d° 1,550 0.13

aQuantitations were doneaccordingtothe procedures described elsewhere (K. Beemon,P. Duesberg, and P.Vogt, Proc.Nat.A-cad. Sci.U.S.A., inpress).

b Anotherpossible spottocorresponding spotcofFig. 7c.

units existwhich contain poly(A) while others

lack poly(A), but appear otherwise identical, could reflect either heterogeneity of the 60 to 70S RNA population extracted from our virus

preparationorsubunit differences within the60

to70S complex. Since binding of RNA, bothto

Millipore filters and to oligo(dT)-cellulose, is thoughttoinvolvepoly(A)sequences,and since over 90% of the 60to 70S RNA binds to these substrates, itappearslikely thatmostofthe30 to40S RNA species without poly(A) arepartof

a 60 to 70S complex which includes other

h

W.

1526

on November 10, 2019 by guest

http://jvi.asm.org/

subunits containing poly(A). If correct, this

would be an independent argument in favor of

the subunitstructure of 60 to70SRNA; it would

imply that 60 to 70S complexes exist between

poly(A)-containing and poly(A)-free 30 to 40S

subunits. The phenomenon could not be

ex-plained interms of a hypothesis which assumes

that the 60 to 70S RNA is a conformational isomer of 30 to 40SRNA.

Results of three different experiments now

suggest that the poly(A) of tumor virus RNA is at or near the 3' end of the RNA. It was shown

bytwodifferent end-group methodsthat

adeno-sine is the 3' terminal nucleotide of poly(A) as wellas ofviralRNA (2, 38, 43, and J. Keith, M.

Gleason, and H. Fraenkel-Conrat, Proc. Nat.

Acad. Sci. U.S.A., in press). Nevertheless,

end-group analyses of macromolecules are subject to error, because only a very small fraction of

the molecule is analyzed. However, these

re-sultsare complementedby our observation that

RNaseHremoves poly(A) from viral RNA

with-out significantly reducing its size. The sum of

these experiments suggests strongly that the

poly(A)of 30 to40Stumor virus RNA is at the 3'

endofthepolynucleotide,aswith other

poly(A)-containing RNAs (31, 32, 40). There is no

evi-dencefor internal poly(A), because (i) one

poly-(A) stretch ofthe size found in this study can

account forallpoly(A) associated withan aver-age 30 to 40S RNA, (ii) 30 to 40S RNAs with

and without poly(A) (eithernaturally occurring

orafter treatment with RNase H) have

approxi-mately the same size, (iii) the poly(A) of 30 to

40S tumor virus RNA was found to reside at a

unique location relative to other

heteropoly-mericsequences of the RNA. However, prelim-inary experiments suggest that A-rich runs,

other than the poly(A) of 200 nucleotides

de-scribed above, exist in the RNA. This

sugges-tion isbasedonthe observation that 20 to 25%

of all 11S fragments of 30 to 40S RNA were

found to bind to oligo(dT)-cellulose, whereas

only 5 to 10% of such fragments would be

expected to bind ifbinding were solely due to

poly(A)-tagged fragments of an RNA with an

originalsize ofaround3.5 x

106

(12).

Some,

perhaps

all, individual RNase

T,-re-sistant oligonucleotides appear to have the

samerelative location onall the 30to40SRNA

molecules with respect to-the

position

of

poly(A), suggesting that individual 30 to 40S

RNAs are not

permuted

with

regard

to se-quences of the RNA. This is

compatible

with

geneticexperimentswhich indicate that certain

markers of different avian tumor virus strains

recombine at stable

frequencies

(J.

Wyke,

J. Bell, and J. Beamand, Cold Spring Harbor

Symp.

Quant. Biol., in press; P. K. Vogt,

per-sonal communication).Thus, thepoly(A) should prove a useful marker for mapping oligonucleo-tides on the RNA.

Given acomplexityof only 3.5 x 106daltons

forRSVRNA (K. Beemon, P. Duesberg, and P. Vogt, Proc. Nat. Acad. Sci. U.S.A., in press) and several (two to three) 30 to 40S subunits

(H. Delius, P.

Duesberg,

and W.

Mangel,

Cold

SpringHarbor

Symp.

Quant.

Biol.,

in

press)

for

60to70S RNA it maybe

argued

that the

differ-ent sequences are

distributed

unevenly

over

several subunits. In this case one subunit may

contain onlyone-halfofall different sequences,

but all ofthesesequenceswould berepresented

twice per 30 to 40S polynucleotide. The

com-plexityofpoly(A)-tagged fragmentsof 30 to40S

RNAwould then notonly bea function of their

size but also of the number of subunits over

whichthe different sequences of RSV RNA are distributed. The lowcomplexityofthe 10 to11S

fragments analyzed here (containing one to

three largeoligonucleotidesout ofabout20 to 30

per 30 to40SRNA) argues against the

possibil-ity that the differentsequences ofRSV RNAare

distributed repetitively over many subunits.

Although our evidence does not exclude the

possibility that sequences of RSV RNA are

unequally distributedovertwo subunits, thisis

considered unlikelyforotherbiologicalreasons.

Nondefective sarcoma viruses containing 30 to

40S RNA of size class a segregate

transforma-tion-defective viruses which contain

only

30to

40S RNAof sizeclass b atratherhigh

frequen-cies(11,29). If the sequences ofRSVRNA were

distributed over several different 30 to 40S

subunits transformation defectives would be

expected which contain both class aandclass b

RNA. This was neverobservedin many

analy-ses of

transformation-defective

viruses (11; P.

Duesberg and P. Vogt,

unpublished

observa-tions).

ACKNOWLEDGMENTS

We thank J. Hurwitz, J. Leis, M. Jacquet, and I. Berkowerforseveralgiftsof RNase Hessentialforthisstudy.

We aregratefultoSunYung Kim and MarieStanleyfor

as-sistance with these experiments, and Karen Beemon, Jan Maisel, andEliCanaaniforreview of themanuscript.

This work was supported by Public Health Service

re-search grant CA-11426 from the National CancerInstitute

andbythe CancerProgram-NationalCancerInstitute,under Contractno.N01CP 43212.

ADDENDUM IN PROOF

It wasshown recentlythat the fingerprintpattern of the transformation-defective deletion mutant(11, 12, 29) of PR RSV-B lacks the two oligonucleotide

on November 10, 2019 by guest

http://jvi.asm.org/

spotstermedaand b in Fig. 7but is otherwise

indis-tinguishable from thatofPR RSV-B RNA(Fig. 7;P. Duesberg and P. Vogt, unpublished observations). Moreover, theRNA ofthis deletionmutant, like that ofother transformation-defective mutants(11, 12, 29) isabout 12 to 15%shorter than that of thewild-type PR RSV-B (P. Duesberg and P. Vogt, unpublished observations). Thus, it appears that RNAsequences

concerned with cell transformation and lacking in this transformation-defective mutant are located

within5to10%ofthe3'terminusof PR RSV-BRNA. LITERATURE CITED

1. Adesnik, M., M.Salditt,W.Thomas, and J. E. Darnell.

1972. Evidence that all messenger RNA molecules (except histone messenger RNA) contain poly(A) se-quencesand that thepoly(A)hasanuclearfunction. J.

Mol. Biol. 71:21-30.

2. Ahmad, M. S.,P. D.Markham, andD. G.Glitz. 1972.

Terminal nucleotides of avian myeloblastosis virus RNA and ofribosomal RNA from chicken leukemic myeloblasts. Biochim. Biophys.Acta281:554-563.

3. Armstrong, J. A., M. Edmonds, H. Nakazoto, B. A.

Phillips, and M. H. Vaughan.1972.Polyadenylic acid

sequencesinthevirion RNA ofpoliovirus andeastern

equine encephalitis virus. Science176:526-528. 4. Aviv, H., and P. Leder.1972.Purificationofbiologically

active globin messenger RNA by chromatography on

oligothymidylic acid-cellulose. Proc. Nat. Acad. Sci. U.S.A.69:1408-1412.

5. Beers, R. F., Jr.1960.Hydrolysisofpolyadenylic acidby pancreaticribonuciease. J. Biol.Chem.235:2393-2398. 6. Canaani, E., and P. Duesberg. 1972. Role ofsubunitsof

60 to 70S avian tumor virus ribonucleic acid in its

template activity for the viral deoxyribonucleic acid polymerase. J. Virol. 10:23-31.

7. Darnell, J. E., R. Wall, and R. J. Tushinski. 1971. An

adenylicacid-richsequenceinmessengerRNAofHeLa

cells and itspossiblerelationshiptoreiteratedsitesin

DNA.Proc. Nat. Acad. Sci. U.S.A.68:1321-1325. 8. Darnell, J. E., L. Philipson, R. Wall, and M. Adesnik.

1971. Polyadenylicacidsequences:rolein conversion of nuclear RNA into messenger RNA. Science

174:507-510.

9. Duesberg, P. H. 1968. Physical propertiesofRous

sar-coma virus RNA. Proc. Nat. Acad. Sci. U.S.A.

60:1511-1518.

10. Duesberg, P. H.,H. L.Robinson, W. S. Robinson,R.J. Huebner, and H. C. Turner. 1968. Proteins ofRous

sarcomavirus.Virology36:73-86.

11. Duesberg, P. H., and P. K. Vogt. 1973. RNA species obtained fromclonallines of avian sarcomaandfrom

avian leukosisvirus.Virology54:207-219.

12. Duesberg, P. H., and P. K. Vogt.1973.Gel electrophore-sis ofavianleukosisandsarcomaviralRNAin

formam-ide: comparison with other viral and cellular RNA species.J. Virol. 12:594-599.

13. Edmonds, M.,and M.G. Caramela. 1969.Theisolation andcharacterizationofadenosine monophosphate-rich polynucleotides synthesized byEhrlich ascites cells. J. Biol.Chem.244:1314-1324.

14. Ehrenfeld, E.,and D. F. Summers.1972.Adenylate-rich

sequencesinvesicular stomatitis virusmessenger ribo-nucleicacid. J.Virol. 10:683-688.

15. Erikson,R.L., E.Erikson,and T. A. Walker. 1971.The

identificationofthe 3-hydroxylnucleoside terminus of avianmyeloblastotisvirusRNA.Virology45:527-528.

16. Gillespie, D., S. Marshall,and R.C. Gallo.1972.RNAof

RNAtumourviruses containspoly(A).Nature N. Biol.

236:227-231.

17. Green, M., and M. Cartas. 1972. Thegenomeof RNA

tumor viruses contains polyadenylic acid sequences.

Proc. Nat.Acad. Sci.U.S.A. 69:791-794.

18. Greenberg, J. R., and R. P. Perry. 1972. Relative occur-renceofpolyadenylic acidsequencesinmessengerand heterogeneous nuclear RNA of L cellsasdeterminedby poly(U)-hydroxylapatite chromatography. J. Mol. Biol.

72:91-98.

19. Horst, J., J. Keith, and H. Fraenkel-Conrat. 1972. Characteristic two-dimensionalpatternsofenzymatic digests ofoncorna and other viral RNAs. Nature N.

Biol. 240:105-109.

20.Ihle, J. N., K-L. Lee,and F. T.Kenney.1974. Fractiona-tion of 34S ribonucleic acid subunits from

oncor-navirusesonpolyuridylate-Sepharose columns. J. Biol.

Chem. 249:38-42.

21. Johnston, R. E., and H. R. Bose. 1972. Anadenylate-rich

segmentin the virionRNA of Sindbis virus. Biochem. Biophys. Res. Commun.46:712-718.

22. Kates, J. 1970. Transcription of the vaccinia virus

ge-nome and the occurrence of polyriboadenylic acid

sequences in messenger RNA. Cold Spring Harbor Symp. Quant. Biol. 35:743-752.

23. Lai, M. M. C., and P. H. Duesberg. 1972. Adenylic acid-richsequencein RNAs of Roussarcomavirusand Rauscher mouse leukaemia virus. Nature (London) 235:383-386.

24. Leis, J. P.,I.Berkower,andJ. Hurwitz.1973.Mechanism

of action ofribonuclease Hisolatedfromavian

myelo-blastosis virus and Escherichia coli. Proc. Nat.Acad. Sci.U.S.A. 70:466-470.

25. Lewandowski, L. J., and S.H.Leppla.1972.Comparison of the3' termini ofdiscretesegmentsof the double-stranded ribonucleicacidgenomesofcytoplasmic poly-hedrosis virus, wound tumor virus, and reovirus. J. Virol. 10:965-968.

26. Lim, L., andE. S.Canellakis. 1970.Adenine-rich

poly-mer associated with rabbit reticulocyte messenger

RNA.Nature (London) 227:710-712.

27. Lindberg, U., andT.Persson. 1972.Isolation of mRNA from KB-cells by affinity chromatography on

polyu-ridylic acid covalently linked to Sepharose. Eur. J.

Biochem.31:246-254.

28. Maitra, U.,Y.Nakata,and J. Hurwitz. 1967.Therole of

deoxyribonucleic acid in ribonucleic acid synthesis. XIV. A study of the initiation of ribonucleic acid synthesis.J. Biol.Chem.242:4908-4918.

29. Martin, G. S., and P. H. Duesberg. 1972.The a-subunit

in the RNA oftransforming avian tumor viruses: I.

Occurrence in differentvirus strains. II. Spontaneous

loss resulting in nontransforming variants. Virology

47:494-497.

30. Maruyama, H. B., M. Hatanaka, and R.V.Gilden.1971. The 3-terminal nucleosides of the high molecular weight RNA of C-type viruses. Proc. Nat. Acad. Sci. U.S.A.68:1999-2001.

31. Mendecki, J., S. Y. Lee, and G. Brawerman. 1972. Characteristics of thepolyadenylicacidsegment associ-ated withmessengerribonucleic acidinmousesarcoma

180ascitescells.Biochemistry11:792-798.

32. Molloy, G. R., M. B. Sporn, D. E. Kelley, and R. P.

Perry. 1972. Localization of polyadenylic acid

se-quences in messengerribonucleic acidofmammalian

cells.Biochemistry11:3256-3260.

33. Pemberton,R.E., and C. Baglioni.1972.Duck hemoglo-binmessengerRNA containsapolynucleotidesequence

rich inadenylicacid. J. Mol. Biol.65:531-535. 34. Penman, S.,M.Rosbash,and M.Penman.1970.

Messen-ger and heterogeneous nuclear RNA in HeLa cells:

differential inhibitionby cordycepin. Proc.Nat. Acad.

1528

on November 10, 2019 by guest

http://jvi.asm.org/

LOCATION OF POLY(A)IN RSV RNA Sci.U.S.A. 67:1878-1885.

35. Philipson, L., R. Wall, G. Glickman, and J. E. Darnell. 1971. Addition of polyadenylate sequences to virus-specificRNA duringadenovirus replication. Proc. Nat. Acad.Sci. U.S.A. 68:2806-2809.

36. Pollet,R.,P.Knolle, andC. Weissmann. 1967. Replica-tion ofviral RNA. XV. Purification and properties of

Q, minus strands. Proc. Nat. Acad. Sci. U.S.A.

58:766-773.

37. Pridgen, C.,and D. W. Kingsbury. 1972.Adenylate-rich

sequences in Sendai virus transcripts from infected

cells.J. Virol. 10:314-317.

38. Rho, H. M., and M. Green.1974.Thehomopolyadenylate andadjacent nucleotidesatthe3-terminus of30-40S RNA subunits in the genome of murine

sarcoma-leukemia virus. Proc. Nat. Acad. Sci. U.S.A.

71:2386-2390.

39. Shelton, K. R., and J. M. Clark, Jr. 1967. A proton

exchangebetweenpurines andwaterand its

applica-tiontobiochemistry. Biochemistry6:2735-2739. 40. Sheldon, R., J. Kates. D.E.Kelley,and R. P. Perry. 1972.

Polyadenylic acidsequences on3' termini of vaccinia

messenger ribonucleic acid and mammalian nuclear

and messenger ribonucleic acid. Biochemistry 11:3829-3834.

41. Spirin,A.S.1963.Someproblems concerning the

macro-molecularstructureofribonucleic acids. Progr. Nucleic Acid Res. 1:301-345.

42. Staynov,D. Z., J. C.Pinder, and W. B. Gratzer. 1972.

Molecularweight determinationof nucleicacids bygel electrophoresis in non-aqueous solution. Nature N.

Biol.235:108-110

43. Stephenson,M.L.,J. F.Scott,and P.C. Zamecnik. 1973.

Evidence thatthe polyadenylic acidof"35S"RNA of avian myeloblastosis virus is located at the 3-OH terminus. Biochem.Biophys. Res.Commun.55:8-16.

44. Wang,L-H., andP. H.Duesberg.1973. DNApolymerase ofmurine sarcoma-leukemia virus: lackofdetectable

RNaseHandlowactivitywith viral RNA and natural

DNAtemplates. J.Virol. 12:1512-1521.

45. Weinberg, R. A.,Z. Ben-Ishai,and J. E. Newbold. 1972.

Poly(A) associatedwithSV40messengerRNA. Nature N.Biol.238:111-113.

46. Wilt,F. H.1969. An artifact in the alkalinehydrolysisof

RNA labeled with guanosine-8-T. Anal. Biochem.

27:186-189.

47. Yogo,Y., andE. Wimmer. 1972.Polyadenylicacidatthe

3-terminus ofpoliovirusRNA. Proc. Nat. Acad.Sci.

U.S.A. 69:1877-1882. VOL.14,

on November 10, 2019 by guest

http://jvi.asm.org/