Sequence of the Long Terminal Repeat and Adjacent Segments of the Endogenous Avian Virus Rous-Associated Virus 0

(1)

0022-538X/82/070191-10$02.00/0

Sequence of

the Long

Terminal

Repeat and Adjacent

Segments

of

the

Endogenous

Avian Virus

Rous-Associated

Virus

0

STEPHEN H. HUGHES

ColdSpring Harbor Laboratory, Cold Spring Harbor, New York 11724 Received 9 December1981/Accepted 9 March 1982

Rous-associatedvirus 0(RAV-0), anendogenouschicken virus, does not cause

disease when inoculated into susceptible domestic chickens. An infectious

unintegrated circularRAV-0 DNA was molecularly cloned, and thesequence of

the long terminal repeat (LTR) and adjacent segments was determined. The

sequenceof the LTRwasfoundto bevery similar to that ofreplication-defective

endogenous virusEV-1.LiketheEV-1 LTR, the RAV-0 LTR is smaller (278 base

pairs instead of 330) than the LTRs ofthe oncogenic members of the avian sarcomavirus-avian leukosis virusgroup. There is,however, significant

homolo-gy. The most striking differences are in the U3 region of the LTR, and in this

region thereare aseries ofsmall segments present in theoncogenicviruses which

areabsent in RAV-0.Thesedifferencesin theU3 regionof the LTRcouldaccount

forthe differences in the oncogenic potential of RAV-0 and the avian leukosis

viruses. I also compared the regions adjacent to the RAV-0 LTR with the

availableaviansarcomavirus sequences. Asegment ofapproximately200bases

tothe rightof the LTR (towardgag)isalmost identical in RAV-0 and thePragueC

strainofRoussarcomavirus. The segmentof RAV-0 which liesbetweenthe end

of theenvgeneandU3is approximately190 bases inlength. Essentiallythisentire

segmentis present between env and src inthe Schmidt-RuppinAstrainofRous

sarcomavirus.Mostofthissegmentisalsopresent betweenenv and src in Prague

C; however, in Prague C there isanapparent deletion of 40 bases in theregion

adjacent to env. In Schmidt-Ruppin A, but not in Prague C, about half of this segment is also present between src and the LTR. This arrangement has

implications for the mechanism by which src was acquired. The region which

encoded the gp37portionofenvappears tobeverysimilar in RAV-0 and theRous

sarcoma viruses. However, differences at the very end ofenv imply that the

carboxy termini of RAV-0, Schmidt-Ruppin A, and Prague C gp37s are

signifi-cantly different. Theimplications of theseobservationsareconsidered.

Rous-associated virus 0(RAV-0), an

endoge-nous chicken virus spontaneously produced by

certainlines ofinbred whiteleghornchickens, is

closely relatedtotheother members oftheavian

sarcomavirus-avianleukosis virus

(ASV-ALV)

group. It has, however, a distinct host range,

defined by the virus-encoded envelope gene,

and,incontrast totheotherASV-ALV viruses,

isnononcogenic eveninlong-term(over1year)

invivo infections of domestic chickens (7, 20).

The ASV-ALV viruses transform cells

by

at

leasttwodistinctmechanisms: certain

oncogen-ic viruses thatcandirectlyandinevitably cause

transformationcarryhomologsofcellulargenes

(oncogenes) (1); the ALV viruses which lack

oncogenes can, at least insomecases,activate

resident cellular oncogenes after

integrating

nearby(14, 21, 22a,23). Asaconsequence, the

ALV viruses are less efficient in causing

neo-plastic diseases than are retroviruses carrying

oncogenes. RAV-0, which does not carry a

cellularoncogene, is apparently also incapable

of activating cellular oncogenes in domestic chickens by insertion since it does not cause

tumors in these birds. There are two

simple

hypotheses: that the RAV-0 proviruses do not

integrate in the same places that ALV

provi-rusesintegrate, i.e., nearthe cellular oncogene c-myc, or that RAV-0provirusescanintegratein

thesamesites, butfailtoactivate c-myc.

What-ever model is correct, there must be some

underlying difference between the

oncogenic

ALVs and RAV-0.

Sincethelongterminal repeat

(LTR)

region

of

viral DNA appears to contain the sequences

necessaryfor theinitiation of

transcription,

ter-mination of RNA and addition of the

polyade-nylatetail, and is

intimately

involvedin

integra-191

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

tion, whichever model proves correct, the crucial difference(s) may reside within the LTR of the virus.

By T1 oligonucleotide fingerprinting (3), hy-bridization (22), and analysis with restriction

enzymes (25), the major differences detected

between theALVs andRAV-0 lie in the 3' end

of the RNA genome, and studies with in vivo

recombinants suggest, but do not prove, that

thisistheregion responsiblefor the difference in

oncogenic potential between the ALVs and

RAV-0 (6, 24, 29). I have cloned unintegrated

circular RAV-0 DNA and shown that the DNA

gives rise toreplicatingRAV-0 virus upon

trans-fection. Ihavesequencedthe LTR of this DNA and compared the RAV-0 sequence with the

sequences of other members of the ASV-ALV

group. Itis nosurprisethat the LTR of RAV-0 is

most closely relatedtothe LTRof the inactive

endogenous virus EV-1, but isrelated, although

lessclosely,tothe LTRsof the other ASV-ALV

viruses. TheRAV-0 LTR issignificantlyshorter

thanthe LTRofRSV,being278(insteadof330)

base pairs in length. The majordifferences

be-tweenRAV-0 and the other

replication-compe-tentASV-ALVviral DNAs lie in theU3segment

ofthe LTR.Thereareminor differences in other

noncodingregions: inU5, the segment

compris-ing the 5' leader, and in the segment upstream

from U3, whichcorrespondsmorecloselytothe

untranslatedregion ofRSVbetweenenvandsrc

than between src and the LTR. There are also

changes in the end of the env gene, the most

obvious of which appear to cause the RAV-0env

genetoterminate downstream of where the env

gene terminatesin RSV.

MATERIALS AND METHODS

Growth of virus; preparation and cloning of uninte-grated RAV-0 DNA. RAV-0 virus was obtained from Harriet Robinson, and a high-titer stock was grown on line15Bchicken embryo fibroblasts. One liter of virus was concentrated approximately30-fold by centrifuga-tion in a Beckman L19 rotor and used to infect two roller bottles of QT6 cells. Seventy-two hours after infection, viral DNA was prepared by the method of Hirt (15) and treated with RNase and pronase. The supercoiled forms of integrated viral DNA were en-riched by a modification of the acid-phenol extraction procedure of Zasloff et al. (9, 30). The recovery of supercoiled viral DNA was monitored at various stagesin the purification by gel electrophoresis, trans-fer to nitrocellulose, and hybridization with 32P-la-beled viral cDNA. Since mapping experiments did not reveal a restriction enzyme suitable for cloning inX that cleaved the viral genome only once (25), the circular viral DNA was partially digested withEcoRI,

which cuts circular RAV-0 DNA twice. This partially digested DNA was ligated to EcoRI-cleaved Charon 3Alac DNA. The resulting A chimeras were packaged in vitro and plated on DP50 SupF. Approximately 0.1% of the plaques contained viral DNA. We have

characterizedadozenclones,oneof which appears to containanentire copy of the viral genome andcanbe rescuedbytransfection of the cloned DNA into chick-encells. The clone(R8) has been mapped with restric-tion endonucleases and the transfer procedure of Southern (27), and it apparently contains a normal copyof theviral genome. R8wascleavedto

comple-tion with HindIlI and EcoRI and subcloned into pBR322. A clone was derived that begins in the

HindlIlsite inenvand ends in theEcoRIsitenearthe end of gag. DNA from this clone was digested to

completion withHindllland treated with Ba131

(Be-thesda Research Laboratories). The ends of the Bal-treated DNAwererepairedwith the Klenowfragment of Poll and ligated to ClaI linkers (Collaborative Research, Inc.). A series of cloneswasderived with

ClaIsites distributed every 200to300 basepairs in the region between Hindlll and the LTR.

Sequence determination.DNAfrom the clones with ClaI linkers insertedwasdigestedwithClaIandSstI

and labeled in thedigestionmix with[o-32P]dCTP(300

Ci/mmol) and the Klenow fragment of PolI. The resultingfragments,labeleduniquelyattheClaIsite, werefractionatedon a1%agarosegel, electroeluted,

purified bychromatography onDEAE-Sephacel,and

sequenced bythechemicalproceduresof Maxam and Gilbert(19).This sequencewasconfirmedby isolating

a HindIlI to SstI fragment from R8, digesting to

completionwithHinfl, labelingtheresulting fragments

with the Klenow fragment of Poll, separating the

labeledstrandsonapolyacrylamide geland sequenc-ing bothseparatedstrands,andbysequencingin both

directions from the BstEII site in theprimer-binding siteadjacenttothe LTR.

RESULTS

Sequence of the RAV-O LTR and adjacent

regions. TheRAV-0LTRis 278 basepairs long,

smaller than that of any other known

replication-competent retrovirus. Figure 1 shows the

se-quenceof theLTRand theadjacentregions. The

RAV-0 LTR has the structural features that

characterize all retrovirus LTRs (H. E.Varmus,

R. Swanstrom, and D. Baltimore, in J. Coffin,

N. Teich, H. E. Varmus, and R. Weiss, ed.,

Molecular Biology ofTumor Viruses: RNA

Tu-mor Viruses, in press). The sequence

TATA-TAA ispresent at a position 23 bases from the

site where RNA was initiated (Fig. 1). The

sequence usually associated with

polyadenyla-tion, AATAAA, is located 24 bases before the

apparent site ofpolyadenylation. A polypurine

tract(AGAGAGGGGA) is located immediately

adjacent to the left end of the LTR. The

se-quence of theprimer-binding site would suggest

that, like Rous sarcoma virus (RSV) (5, 12),

RAV-0uses a tRNA trp primer. The ends of the

RAV-0 LTR form a small imperfect inverted

repeat (six of the first seven and seven of the

first nine base pairs matched; see Fig. 1). In

addition to these general structural features

present in allretrovirus LTRs, the RAV-0 LTR

showsconsiderable sequence homology with the

http://jvi.asm.org/

(3)

VOL. 43, 1982

GCTTGTAGTT

_CAACATCA

_A

TT~GTPTEEAG

_AACACAGGTC

C I AC G

A

I

T

GG+GC

HinfI

E

I48

E

GCA

BstEII PMB

EV-1 sequence

1CGC

NGU

S8tI

C ST9G

SEQUENCES IN

RAV-O

193

t

HEHHH2CCC ~HCGCTC4ATI

HinfI

TIGTAG

C~GEG

TTAGIGA G T

EI

AAC TAICCC AA CA

I~~~G

'TCG~~~~~~

0TATATCTG A

) ~~~u5beginls

ATTCG IcCC A T4 GTCCA

NTG

CGIC4XC T AT

U H

C

H

T

endsI

GGGXGVIGG8

AEIGCGX;T4

999AC9%999G AT99%99999

GACTTTCCTTG

IcGI

EIA

cc

CGCACCG9IGCGP

~T~9TT%9~

GAAAAGCG4TT

I~E~

AATT9AST9T

GTCCG

AT9G4G

HinfI

ITTAA G|CTA TAG

CA

C

A

9~9:AkIC:k9

GCAAGACAT

_CGTTCTGTAA

T

CGA

c c

cATT

ATCT C CGT CTCA

H~AccI

C

CCGCG

9CCTGCT9

GCTTACGTGTCT

EV-1 sequenoe

lbegins

0 TCTCCG

4 CGC

CCATATGCTC

_GTAT

_{AC GA}

Rbegins

99999

9ISCG

c

ie

I GAI

TA

IT T

T 4

LR9BIA

CTGC Ia98 8 M

FIG. 1. Sequence of the RAV-O LTR and adjoining regions. The sequence of the RAV-O LTR and the

adjacent regionswasdeterminedbythemethods ofMaxamandGilbert(19).Aseriesof subcloneswasprepared

by inserting ClaI linkers (vertical arrows) afterdigestion with BaIl3 exonuclease for various times. These subclonesprovidedadditionalsequencing startsites in theregionsin and around theRAV-OLTR. Prominent structuralfeatures suchasthebeginningand endof theU3, R,andU5regions,thepolypurine tract(PPT),the

primer-bindingsite(PBS),and the endof theenvgenearemarkedonthe sequence(theTAAcodonatthe end of envisboxed).The inverted repeatsattheendsof theLTRaremarkedbyarrowsunderthe sequence. Differences between theRAV-O sequence and thepublishedsequenceof EV-1 (5)areshownasdifferentbasesgivenabove theRAV-Osequence.Theavailable EV-1 sequence is shorter than the sequenceIhavederived forRAV-O,and theextentof the EV-1 sequence is denotedbyapairof bars above the RAV-O sequence. An A-Tpairpresentin theRAV-Osequence and absent in EV-1is denotedbyanopencircle.

http://jvi.asm.org/

[image:3.504.60.453.71.567.2]

(4)

RAV-0O

E

SRA G

INTXC&AT

A

S

MA

I GI T

&&

TT

RU

C AIT&T

TA

EMU4I

WMW

EPT U3 begins

MMERI

IMPH

E2GEE

M

M Ic

MI M?

c444

Mc&

EM

Rbeginis

EWPIM

HUNME

_ _

&TTAT

&C&&IIACT&

&&CCMTC

IC422

RM

GGICCEIC

ACC&CCT&

GCA12T&IGCATE AWMEA4M4

U.beginls

ATT&CTA IC&

ICT&T&HT

UUMHEAT&&

I

MENEM~

MENEM

I

FIG. 2. Comparison of the RAV-O LTR and the LTR of the SR-A strain of RSV (28). The LTR of RAV-O is significantly smaller than the LTR of the other members of the ASV-ALV group. The differences reside principally in the U3 region. I have aligned the U3 regions wherever it was possible, leaving gaps in the RAV-O sequences at positions where the sequences present in SR-A are obviously absent inRAV-O. In some small regions there are no obvious homologies. These regions are indicated by small bars between the RAV-O and SR-Asequences. The distribution of barsbetween the sequences and of gaps within the sequences shows directly the disposition of the differences in the two sequences. The more modest changes inU5and R are alsoindicated in this same fashion. This complete sequence of theRAV-OUsisin perfectagreement with a partial sequence of

RAV-O"strong stop" DNA(Swanstrom, unpublished data). PPTdesignates thepolypurine tract.

LTRs of the other members of the ASV-ALV the LTR ofRAV-O and the LTR of EV-1. Two

group. TheRAV-O LTRis strikingly similar to are transitions, one is a transversion, and the

the LTRof EV-1 (16), adefective endogenous fourth is an additional A-T base pair in the

RAV-virus, and is less closely related to the replica- 0 sequence (Fig. 1). This additional A-Tpair,

tion-competent members of the ASV-ALV near the left end ofU3, is part of a sequence

group. Thereare onlyfour differences between which gave, when the sequence was derived in

1102

PUMP

HUI IM

2402M

.

MEE 14

14 EE21442

MICIMIC2

4 4 . 2 2 . . 2 2

E

11 MR

E

IIE MW

MMU

.HHHM

'f

AA

T.i

HICUP

I 4

4 E

E I

E E

.

I

UUIOEAU

UIAMME

INMEW

.NHEHH

HMMH

MH

.HE

http://jvi.asm.org/

[image:4.504.94.386.77.498.2]

(5)

one direction, a serious compression artifact, RAV-0 LTR, or simple mutations. The

T,

oligo-and I initially misread the sequence. Since this nucleotide used by Tsichlis and Coffin (29) to

misreading of the RAV-0 sequence gave a se- distinguish the endogenous and exogenous

vi-quence identical to that published for EV-1, it is ruses(Cx and

Cn,

respectively) derive from the

possible that EV-1 and RAV-0 are identical in region of U3 immediately adjacent to R. The

this region and that there is an error in the relevant sequences were

GCCACCATCAA-published sequence of EV-1 (16). The GCCACCATCAA-published TAACG in RAV-0, which gave rise to the

T,

sequence of EV-1 includes about 60 base pairs to oligonucleotide 08, and GATACAATAACG in

the right of U5; these are identical in RAV-0. Schmidt-Ruppin A (SR-A), which gave rise to

About 25 base pairs of sequence from the left of the

T,

oligonucleotide C (see Fig. 2).

the EV-1 U3 have been published; there are two There is a large open reading frame ending 165

changes from RAV-0, both transitions (Fig. 1). bases upstream of the LTR that is likely to

Thesimplest explanation of this similarity is that encode env. Not only is this the only sizable

EV-1, and the other endogenous chicken virus- open reading frame (see Fig. 4), but there is also

es, arose by germ line infection of a virus very considerable homology with the proposed env

closely related to RAV-0 (10, 17, 18, 25). gene of the SR-A (8) and Prague C (Pr-C) (D.

Because the sequences of the EV-1 LTR and Schwartz, personal communication) strains of

the RAV-0 LTR are nearly identical, the rela- RSV. When these sequences are compared in

tionship between the RAV-0 LTR and the RSV the region which corresponds to the carboxy

LTR is very similar to the relationship between terminus of env, a divergence is found. The Pr-C

the LTRs of RSV and EV-1 (16). The RAV-0 env gene is longer than that of SR-A, and the

LTRis 55 base pairs shorter than the RSV LTR, sequence of the RAV-0 env gene, which is closer

and the sequences could be aligned as though to that of SR-A than to that of Pr-C, is the

the RAV-0 LTR had suffered numerous small longest of the three. The inferred amino acid

deletions or, conversely, as though the RSV sequence of these three env genes is given in

LTR had acquired numerous small insertions Fig. 3; that of the DNA is given in Fig. 4.

(Fig. 2). Even when the two sequences are We have sequenced the region between env

aligned optimally, there are still small regions of and src in SR-A and found significant differences

nonhomology.These could be explained as dele- between our data (J. Sorge and S. Hughes,

tions from the RSV LTR, insertions into the unpublished data) and those published by

Czer-RAV- IEL GLY VAL SER

ASPRPD

ILEU GLY

SERTR

E

ARiG

SRAR eV" It of R .E1 itI of of It It

GLY Wi PHE GLY GLY ILEU GLY

GLUW

TI ALA VAL HIS LEJ LEU

It LEJ

LEUJ

It It It it It to It N N N It It M It R It R It o

it

It of t

L.EU N N N'~ N

LYS GLY LEIJ LE LEJ GLY LEU VAL VAL ILEJ LEU LEJ LEJ VAL

if if i It if of i it NIt I I N R It IN It I NIIt it It o Nf I I N It It

VAL CYS LEUJ PRI) CYS LUE LE GLN IWIE VAL SER SER SER ILEI

i n to It of NI t of PHE

nI

N t it N

f i n II N ItN N M

ET

EU CYS GLY ASN ARG

ARi LYIsET ILE ASN AS ERSE ER

TYR

HIS THR

GLU

YR

n N aN Rtt SER ASN of to N R

NIf

N n N

ASNN

GLU LYS

PRO

HIS

GLY

ILEJ

LYS LYS lEU GlN LYS AU CYS ARG

GIN

PRO GLU ASN GLY ALA VAL

[image:5.504.114.398.400.634.2]

Ap3

N MET " GLY GLY ALA VAL

FIG. 3. Inferred amino acid sequences of the ends of the env genes ofSR-A,Pr-C,andRAV-O.Theimplied amino acid sequences of the env genes of SR-A (8), Pr-C (Schwartz,personalcommunication),andRAV-Ohave beenaligned by maximizing homology at the nucleic acid and protein levels. The nucleic acid sequences from the region at the end of env are compared in Fig. 4.

VOL.43,1982

SEQUENCES

IN RAV-0 195

http://jvi.asm.org/

(6)

196

_SRA

HUGESITEA

J.129VS

ViSSR4A888V

M

RoL.

_

_%99I%

_ _ _ _

PRC

TCGGAACGGXT

TI4G4ATIII

ICG4I

4 T4AC4G

4CI4I

Sequence ofSR-A not checked *Sequence of SR-A checked

Cl;CC

ACGA

CAATCGCTCGTG

CTA AT

TCT

TCAGTAGCGcI

T

TU

G

~

C AG

H

IG TT T C

A

_~~~~~~~~~

C

TAT ACcTAT A A

T(

A

TCG

AT

GG~~GC44TTA

C4CGGCCTT~~

TACGAT

TGGGGG

ITATCGICTA --- ---

--I~

A

T

G

A

C e C I I G C C T

IrGC

C C C I I GC9I

4 CG

C

T

ITIC

TTT

GIICGC4

A

T

G

CGCCGICTA

T

CCGCCGGGATTCGG

T

G4CI(

A

G

GGCTGCIATIGCG TIcIT I 444CGCC

C

IAc I

EGC

CT

[image:6.504.59.457.66.483.2]

_ _ _ __

FIG. 4. Comparisonof the sequences fromenvtothe LTR of RAV-O with sequencesfromenvtowardsrcin SR-A and Pr-C. TheregionofRAV-0 from the end ofenvtothe LTRwasalignedwith theendofenvandthe

in-tergenic regionbetweenenvandsrcof bothPr-C(Schwartz,personalcommunication)andSR-A(8). Wehave

checked a portion of the SR-A sequence in thisregion, using subclones derivedfrom the same X clone (9)

sequenced byCzernilofskyetal.(8)and foundsignificantdifferences in theregionbetweenenvandsrc(J.Sorge

andS. Hughes,unpublisheddata).TheSR-A sequencegivenin thefigureisacompositeof thepublisheddata

andourunpublishedrevision. Theportionnotchecked is indicated in thefigure, tothe leftof thearrowhead;

thosethathavebeen checkedareto therightof the arrowhead.Sinceaportionof thesequence foundinRAV-O

and SR-A near the end of the env gene is apparently deleted in the Pr-C sequence (Schwartz, personal

communication),Ihave indicatedthisapparentdeletionbyagapin thePr-C sequencefilled inbyadotted line. Smaller deletions(usuallyasinglebasepair) areshownasgaps. Lack ofhomologybetweenSR-A and RAV-0

andPr-C andRAV-o are indicated by barsbetween theRAV-o andSR-A andthe RAV-0andPr-C sequences.

nilofsky et al. (8) (Fig. 4). Although we did not In the case ofSR-A, the region ofhomology

sequenceenvinSR-A,itispossiblethatsomeof continuesbeyond envand extendsfor about 150

the differences between the SR-A and RAV-0 nucleotides, which are

apparently

noncoding,

sequences aredue to errors. Such errorscannot,

diverging

nearthepointinthe RAV-0sequence

however, accountforthe

significant

differences where the

polypur

netractadjacenttothe LTR

between the Pr-C and RAV-0 env

sequences

begins

and in SR-Aabout

halfway

between env

(Fig. 4). and src

(Fig.

4). Asimilar

comparison

withPr-C

J.VIROL.

196 HUGHES

http://jvi.asm.org/

(7)

VOL4319 V-'19

RAV--O

CTATT

'AGT

TG5 CT CT

G

T

G G

A TT A G

PRC

T

A

TAA ECTC

A A

C

GGA

GA

C

A C

CTAATCCC

G

c~

~~~~~G

G

TTAIGC GAT T GGGC GGCCCGCCT T G

TCATACAT ccCGCTTTccG CCGCcCAACCG

IACATAAG

GCC IIA

TCXC G GC

TA CG GC

A

T~GT9EG%GG

TT

~GTEE~

TATATTCCGTA C CGA AGCTGCCGA A CAT G

PPT

TCI4CIH;IGH

PCiAAGT

ACCATCCC

FIG. 5. Comparison of the sequences between env and the LTR of RAV-O with the sequences between src and the LTR of Pr-C andSR-A. The regions upstream of the LTR are compared for the three viruses. Lack of homology is indicated as a bar between the sequences being compared. The polypurine tract is marked (PPT). Thedivergence of Pr-C from the other two viruses is clearly evident by an examination of the frequency and the extent of the barred region between the RAV-0 and Pr-Csequences.

revealsanapparentdeletion,extending from the end of env for about 40 bases. Beyond this

"deletion" thehomologyresumes,and there isa

polypurinetractinPr-C betweenenvandsrcata

site corresponding in relative position to the

polypurinetractin RAV-0 whichliesadjacentto

the LTR.

It is also possible to comparethe sequences

which lie betweensrcandthe LTR of the RSVs

with the region between env and the LTR of

RAV-0. Again,RAV-0 ismorecloselyrelatedto

SR-A than toPr-C (Fig. 5). There is inSR-A a

large direct repeat which flanks src (8). This repeatwasfoundtobehomologoustopartofthe

region between env and the LTR of RAV-0. A

directcomparison of thesequencesbetweensrc

andtheLTR ofSR-A with the regionbetween

envand the LTR of RAV-0 reveals aregion of

homology beginning at the LTR of about 120

base pairs in length; the last 40 base pairs

(approximately)adjacenttoenvin RAV-0 donot

haveaclearhomologin theregionbetweensrc

and the LTR of SR-A. In Pr-C, the region of

homology is muchshorter; only about the first

20base pairs adjacent to the LTR are found in theregionbetweensrcand the LTRof Pr-C. On

the other side of the LTR, the untranslated

leaderto therightofU5isvery

similar

in RAV-0

and Pr-C,althoughthere aredifferences(Fig. 6).

Itispossiblethatthisregionisimportantasthe

recognitionsignalforpackagingRNAinto

viri-ons (26) and that the conservation of the

se-quence inthisregionreflects this requirement.

DISCUSSION

Although RAV-0 is an endogenous

virus,

it

replicatesreasonably well,and thesmallRAV-0

LTR thus contains all of the information

re-quired in a retrovirus LTR. Since the RAV-0

LTR is very similar to the LTR of EV-1, a

defectiveendogenousvirus whichistranscribed

atverylowlevels(13),it is

unlikely

thatthe poor

transcription ofEV-1 is due to a defect in the

EV-1 LTR.The fewchangesbetweenthe EV-1

and RAV-0 LTRs are not in

regions

like the

"TATAA" box (2), which wecan

recognize

as

being involvedin

transcription.

In

addition,

EV-VOL.43, 1982

SEQUENCES

IN RAV-0 197

http://jvi.asm.org/

[image:7.504.132.373.69.353.2]

(8)

U5 PBS

PRC

HRACGC214A

I

CTGG

CTCG§ 2

T0RHUI

A

242qa%%

G

C

[image:8.504.137.376.77.384.2]

T0HH~P

I

A

242248%%

G

GT

FIG.~~~~~~~

6.Coprio of th

sqeCe

of th

uTrasae

UG

lede seune betee

CGIATGA

and gag in Pr- and

4RSE2UH4E

HEHEUEE

HHH

RAV-O. Most of the regions between Us and gag in Pr-C and RAV-O are shown for comparison. Lack of homology is shown by a bar between the sequences.

1 becomestranscriptionallyactive after azacyti-dinetreatment,whichsuggeststhat,under

prop-ercircumstances,the EV-1 LTRcanfunction to

initiatetranscription (11). These data taken

to-gether with earlier studies suggest either that

there maybe contextual constraints on

expres-sion ofproviruses (4)orthat DNAmodification mayplayamajor roleinregulation ofexpression (11), orboth.

Eventhough the RAV-O isacompetent repli-cating virus, it doesnot,despiteaclear homolo-gywith the ALV viruses, induceanyneoplasia

indomesticchickens. Previous studies

suggest-ed, but didnotprove,thatthecrucial difference

inoncogenic potential resides within the LTR (6, 24, 29). Both hybridization andT1 oligonucleo-tide fingerprinting experiments show that the

major differences between the ALVs andRAV-O

areinthe regionatthe 3' endof viralRNA,and

my sequenceanalysishasdemonstrated that the

differencesliemostlywithin the U3 region. The ALVscaninduce leukosis by integratingnextto

the cellular oncogene c-myc and enhancing

c-myctranscription (14, 21, 22a, 23).

RAV-O does not induce leukosis in domestic

chickens; therefore, at some level it does not

duplicateanevent orprocessproduced byALV.

Thereareseveralpossibilities, some morelikely

than others. It is possible the RAV-O cannot

infect theappropriatetargetcell,althoughthis is

unlikely since subgroup E ALVsareoncogenic

(6, 24). It is alsopossible,though unlikely,that

RAV-Oactivatesc-mycbutkills thetargetcell.It

seems more likely either that RAV-O does not

integrate near c-myc orthat RAV-O proviruses

integratednear c-mycdo not enhance

transcrip-tion to a level sufficient for transformation.

Differences intheU3region of the LTRs of ALV

andRAV-Ocould alter the"specificity"of

inte-gration, ifthere is any specificity in retrovirus

integration.Alternatively,the signals for

regula-tionofRNA transcription that reside in the U3

region of the LTR could be responsible for

changes in the ability of RAV-O to activate

adjacentgenessuch as c-myc. Comparing the U3

regions ofALV and RAV-O does not resolve

theseissues, although it is temptingtospeculate

thatwhatever theRAV-OLTRlacks in terms of

http://jvi.asm.org/

(9)

199

sequence may be responsible forthe

concomi-tantloss of oncogenic potential.The fact that the

differences between the ALV and RAV-0 U3s

aredistributedover thewhole U3 region makes a

direct testof this hypothesis difficult.

Do host factors influence the oncogenic

po-tential of the avian viruses, differentiating

be-tween theoncogenic ALVs andRAV-0?A

pre-liminary reportfromWeiss and Frisby suggests

that hostfactorsmayplayakeyrole;they have

found RAV-0 to be tumorigenic in Sonnerat's

jungle fowl (R. Weiss and D. Frisby, in D. F.

Yolm, ed., 10th International Symposium for

Comparative ResearchonLeukemiaand

Relat-ed Diseases,inpress). Althoughthesample size

was small, about half of the RAV-0-infected

birds died ofapparent lymphoid leukosis, and

several more died from "wasting disease."

Even in Sonnerat's jungle fowl, however, the

RAV-0-inducedtumorsappeared on average

sig-nificantly later, and at a lowerfrequency, than

tumorsinducedby the subgroup E recombinant

RAV-60.

Itis not yet clear why RAV-0 is oncogenic in

these particular birdsand not in ordinary

chick-ens, but Sonnerat's jungle fowl completely lack

RAV-0-relatedendogenous viruses, and RAV-0

replicates to a very high titer in these birds.

Whetherone ofthese factors, or others, as yet

undiscovered, distinguishes the oncogenic

sus-ceptibility of chickens and Sonnerat's jungle

fowl, it is clear the differences between the

hosts, as wellas differences between the

virus-es, canprofoundly alter the incidenceof

tumori-genesis. This implies that ahostfactor, or

fac-tors, is involved. It would be particularly

interestingtolearn whether theRAV-0-induced

tumorspresentinSonnerat's jungle fowlarethe

resultofc-mycactivation by RAV-0.

Aside from the LTR sequences, the most

obvious differences between the RSVs which

have been sequenced and theregionof RAV-0

DNAIhavesequencedare attheveryendof the

envgene.If thesequencesattheendsof theenv

genesof Pr-C (Schwartz,

personal

communica-tion), SR-A(8), and RAV-0are

compared,

there

isaclearhomologythatbreaks downatthevery

end of theenvgene.All three envgenes appear

to terminate in different places,

giving

primary

translation products whichare verydifferent at

theircarboxy termini. Thedramatic differences

in predicted amino acid sequence suggeststhat

the extremecarboxyendofgp37is

probably

not functional.

I wassurprisedtofindacomplete

homolog

of theregion betweenenv and the LTRof RAV-0

(excluding the polypurine

tract)

in the

region

betweenenvandsrcofthe SR-Astrain of ASV.

Beyondtheapparentdeletionatthe end ofenv,

Pr-C alsocontainsverysimilarsequences.In

Pr-Cthere isevenapolypurine tract present atthe

same position relative to env as the polypurine

tract which is adjacent to the LTR inRAV-0. If

we assume that RAV-0 never contained src,

then this homology presumably is the result of theway src was acquired by the ancestor ofthe

RSVviruses. This antecedent virus presumably

hadconsiderablehomologytoRAV-0,although,

unless there were additional recombination

events, it hada U3region similarto thatofthe

oncogenicmembers of theALV-ASV group.

Itis unlikelythat, when src was first acquired,

it was simply inserted into the untranslated

region betweenenvand the LTR,sinceinSR-A

the src gene is flanked by an essentially

com-plete copy of sequence which in RAV-0 lies

betweenenv and the LTR. Theidea thatsrcwas

notmerely inserted intotheregion betweenenv

and the LTR is reinforced by the presence of

part ofU3 upstreamofsrc in thePr-C and B77

strains of RSV (Mardon, personal

communica-tion;Swanstrom,personal communication). Itis

likely thattwoviral genomes together with

cellu-lar srcwere required to create SR-A. Through

integration or the actions of reverse

transcrip-taseorboth,c-srcwas apparently added to the

end ofaviralgenomewiththeconcomitantloss

of U3. To complete the virus,asecondcopyof

thesequencefound betweenenvand the LTR of

RAV-0wasthenadded to the end of src together

withU3;almostcertainly a second viral genome

wasthe donor of this segment.

The region between the end of U5 and the

beginningofgag appears tobehighly conserved

between Pr-C RSV and RAV-0. The

primer-binding site is identical, which suggests that

RAV-0,like ASV,uses tRNA trp as aprimerfor

DNAsynthesis. The conservation of therestof

this region argues that there is some functional

constraintonwhat is believedtobeanoncoding

region. Experiments witha mutantvirus witha

150-base-pair deletion in this region imply that

theregionisimportantfor efficientpackagingof

RNAinto virions (26), alikely explanation for

theconservation ofthese sequences.

ACKNOWLEDGMENTS

IthankAnn Mutschler for experttechnical assistance and mycolleagues,especiallyJ.Sorge,G.Payne,and J.Fiddes, forhelpful discussions.

This research was supported byaPublic Health Service grant from theNationalInstitutesof Health.

ADDENDUM IN PROOF

Additionalsequencingin theregionof SR-A which encodes thecarboxyterminus ofgp37does notagree withthe datapublishedbyCzernilofskyetal.(8).The newDNA sequence(J. Sorge, unpublisheddata) pre-dictsthatthecarboxyterminus of SR-Agp37isquite similartothecarboxyterminus ofRAV-0 gp37.

http://jvi.asm.org/

(10)

LITERATURE CITED

1. Bishop, J. M. 1981. Enemies within: the genesisof retro-virus oncogenes. Cell 23:5-6.

2. Breathnach, R., and P. Chambon. 1981.Organization and expression of eucaryotic split genes coding forproteins. Annu. Rev.Biochem. 50:349-383.

3. Coffin, J., M. Champion, and F. Chabot. 1978.Nucleotide sequencerelationships between the genomesofan endog-enous and an exogenous avian tumor virus. J. Virol. 28:972-991.

4. Cooper, G. M., and H. M. Temin. 1976. Lack of infectivity of the endogenous avianleukosisvirus-related genes in the DNAof uninfectedchicken cells. J. Virol. 17:422-430. 5. Cordell-Stewart, B., J. M. Taylor, W. Rhode, H. M.

Goodman, and J. M. Bishop. 1974. Transfer RNAs of Rous sarcomavirus and theinitiationof DNAsynthesis by viral reversetranscriptase, p. 117-135. In A. Kaplan (ed.), Viraltransformationandendogenous viruses, Pro-ceedingsof the 1974 Abraham FlexnerSymposium. Aca-demic Press, Inc., New York.

6. Crittenden, L. B., W. S. Hayward, H. Hanafusa, and A.M. Fadly.1980.Inductionof neoplasmsby subgroup E recombinants of exogenous and endogenous avian

retro-viruses(Rous-associatedvirus type 60). J. Virol. 33:915-919.

7. Crittenden,L.B., R.L.Witter,and A. M.Fadly.1979. Lowincidence oflymphoidtumorsin chickens continu-ously producingendogenous virus. Avian Dis. 23:646-653.

8. Czernilofsky,A.P., A. D. Levinson, H. E.Varmus, J. M. Bishop,E.Tischer, and H. M.Goodman. 1980.Nucleotide sequenceofanavian sarcomavirus oncogene (src)and proposedamino acid sequence for geneproduct. Nature (London)287:198-203.

9. DeLorbe,W., P.Luciw,H. M.Goodman, H.E.Varmus, andJ.M.Bishop. 1980.Molecularcloningand character-ization ofavian sarcoma virus circular DNA molecules. J. Virol. 36:50-61.

10. Frisby, D., R. A. Weiss,M.Roussel, and D.Stehelin. 1979. The distribution of endogenous chicken retrovirus

se-quencesin the DNAofgalliform birds does not coincide with avian phylogenetic relationships. Cell 17:623-634.

11. Groudine, M., R. Eisenman, and H. Weintraub. 1981. Chromatin structureof endogenousretroviral genes and activation of an inhibitorof DNA methylation. Nature (London) 292:311-317.

12. Harada, F., R. C. Sawyer, and J. C. Dahlberg. 1975. A primer ribonucleic acid for initiation of in vitro Rous sarcomavirus deoxyribonucleic acid synthesis. J. Biol. Chem.250:3487-3497.

13. Hayward, W. S., S.B.Braverman, and S. M. Astrin. 1980. Transcriptionalproducts and DNA structure of endoge-nousavian proviruses. Cold Spring Harbor Symp. Quant. Biol. 44:1111-1121.

14. Hayward, W. S., B. G. Neel, and S.Astrin. 1981. Activa-tionof a cellular onc gene by promoter insertion in ALV-induced lymphoid leukosis. Nature (London) 290:475-480.

15. Hirt, B.1967.Selectiveextractionofpolyoma DNA from infected mouse cell cultures. J. Mol. Biol. 26:365-369.

16. Hishinuma,F., P. J. DeBona, S. Astrin,and A. M. Skalka. 1981.Nucleotide sequence of the acceptor site and termi-ni of the integrated avian endogenous provirus ev-J:

integration creates a 6 bp repeat of host DNA. Cell 23:155-165.

17. Hughes, S. H., F. Payvar, D.Spector, G. Payne,R. T. Schimke, H. Robinson,J.M.Bishop, and H. E. Varmus. 1979. Heterogeneityofgeneticloci in chickens:analysis of endogenous viraland non-viral genes by cleavageof DNAwithrestriction endonucleases.Cell 18:347-359. 18. Hughes, S., K. Toyoshima, J. M. Bishop, and H. E.

Varmus.1981.Organizationof theendogenousproviruses of chickens:implicationsfororiginandexpression. Virol-ogy108:189-207.

19. Maxam, A., andW.Gilbert. 1980.Sequencingend-labeled DNA with bare-specific chemical cleavages. Methods Enzymol. 65:499-560.

20. Motta, J. V., L.B. Crittenden, H. G.Purchase, U. A. Stone,W.Okazaki, and R. L. Witter. 1975. Low oncogen-icpotentialof avianendogenousRNAtumorvirus infec-tion orexpression.J.Natl. Cancer Inst. 55:685-689. 21. Neel, B. G.,W.S.Hayward, H. C.Robinson,J.Fang,and

S. M. Astrin. 1981. Avianleukosis virus-inducedtumors

have common proviral integration sites and synthesize discretenewRNAs: oncogenesis by promoterinsertion. Cell 23:323-334.

22. Neiman, P., S. Das, D. MacDonnel, and C. McMillin-Helsel. 1977. Organization of the shared and unshared sequences in the genomes of chicken endogenous and

sarcomaviruses. Cell 11:321-329.

22a.Payne, G. S., J. M. Bishop, and H. E. Varmus. 1982. Multiplearrangements ofviral DNA andanactivated host oncogene(c-myc)inbursallymphomas.Nature(London) 295:209-214.

23. Payne,G.S., S.A.Courtneidge,L.B.Crittenden,A.M. Fadly,J. M.Bishop, and H.E. Varmus.1981.Analysisof avian leukosis virus DNA and RNA in bursaltumors:viral geneexpression is not requiredfor maintenance of the

tumorstate.Cell 23:311-322.

24. Robinson, H. L.,M. N.Pearson,D. W. DeSimoneP.N. Tsichils,andJ.M.Coffin. 1980.Subgroup-E avian-leuko-sis-virus-associateddiseasein chickens. ColdSpring Har-borSymp.Quant.Biol. 44:1133-1142.

25. Shank, P., S. Hughes, and H. E. Varmus. 1981.Restriction endonuclease mapping of the DNA of Rous associated virus 0 reveals extensive homology in structure and sequence with avian sarcoma virus DNA. Virology 108:177-188.

26. Shank, P. R., and M. Linial. 1980. Avian oncovirus mutant(Se2lQlb)deficient ingenomic RNA: character-ization ofadeletionin theprovirus.J.Virology 36:450-456.

27. Southern, E. M. 1975. Detection of specific sequences amongDNA fragments separated by gel electrophoresis. J. Mol.Biol. 98:503-517.

28. Swanstrom, R., W. DeLorbe, J. M. Bishop, and H. E. Varmus. 1981. Sequencing of cloned unintegrated avian

sarcomavirus DNA reveals direct and inverted repeats similar to those in transposable elements. Proc. Natl. Acad.Sci. U.S.A.78:124-128.

29. Tsichlis,P. N.,and J. M. Coffin. 1980. Role of the C region in relative growth rates of endogenous and exogenous avian oncoviruses. Cold Spring Harbor Symp. Quant. Biol. 44:1123-1132.

30. Zasloff, M., G.D.Ginder, and G. Felsenfeld. 1978. A new methodfor the purification of covalently closed circular DNA molecules. Nucleic Acids Res.5:1139-1152.

http://jvi.asm.org/