Molecular cloning and characterization of avian sarcoma virus UR2 and comparison of its transforming sequence with those of other avian sarcoma viruses.

(1)

Vol.50,No. 3 JOURNALOFVIROLOGY, June 1984,p. 914-921

0022-538X/84/060914-08$02.00/0

Molecular Cloning and Characterization of Avian Sarcoma Virus

UR2

and

Comparison of

Its

Transforming Sequence

with

Those

of

Other Avian

Sarcoma

Viruses

WENDIS. NECKAMEYER ANDLU-HAIWANG*

TheRockefeller University, New York, New York 10021

Received 3 January 1984/Accepted13 March1984

Avian sarcoma virus UR2 and its associated helper virus, UR2AV, were molecularly cloned into AgtWES * XB by using unintegratedviral DNAs. One UR2 and severalUR2AVcloneswereobtained.The

UR2 DNA was subsequently cloned into pBR322. Both

UR2

and UR2AV DNAs were tested for their

biological activity by transfection onto chicken embryo fibroblasts. When cotransfected with UR2AV

DNA, UR2 DNA was able to induce transformation of chicken embryo fibroblasts with a morphology

similar to that of parental

UR2.

UR2-specific protein with kinase activity and UR2-specific RNA were

detected in the transfected cells. Transforming virus,

UR2(UR2AV),

was produced from the doubly

trantsfected

cells.Five of the sixUR2AVclonestested were alsoshowntobe biologicallyactive. The insert

of the UR2 DNA clone is 3.4 kilobases in length and contains two copies of the

long

terminal repeat.

Detailed restriction mapping showed that UR2 DNA shared with UR2AV DNA 0.8 kilobases of 5' sequence, including a portion of 5' gag, and 1.4 kilobases of 3' sequence, including a portion of 3' env. The

UR2

transformingsequence, ros,is ca. 1.2kilobases.No significanthomology was found between v-ros and theconserved regions of v-src,v-yes, or v-abl.Bycontrast, asignificanthomology wasfoundbetween v-ros andv-fps. The v-fps-related sequence was mapped within a 300-base-pair sequence in the

middle

of ros.

Avian sarcoma virus (ASV) UR2 isa recently character-ized replication-defective virus, which is able to induce sarcomas inchickens and efficientlytransform chicken em-bryo fibroblasts (CEF)in culture (2). CEF transformed by

UR2arecharacterizedbyanextremely elongated morpholo-gy(2). Previous analysis of the UR2 RNA genome showed that it shares with its associated helper virus, UR2AV, ca.

2.1kilobases(kb) of 5' and 3' sequences and containsatthe middle of thegenome ca. 1.2 kb ofspecificsequence, called ros (31). Studies of the ros sequence by hybridization and oligonucleotide fingerprintingshowed that it is distinct from the transforming genes of other known ASVs and acute leukemia viruses(21, 31). The normalcellular DNAhomolog of the ros sequence has been detected in chickens, quails, and ducks(21). However, the expressionofc-rosinvarious tissues andorgansof 10- to14-day-oldchickens isverylow (less than 1copy percell), exceptinthekidney (2.5 copies

percell) (21).

The UR2-infected cells produced only the 24S genomic RNA(31), which wasshown to encode fora68,000-dalton gag-ros fusion protein, P68, that was associated with a tyrosine-specific protein kinase activity (7). Theenzymatic properties of P68 are distinctive from those of other ASV protein kinases, i.e., Rous sarcomavirus(RSV) p60,

Fujin-ami sarcoma virus P140, and Y73 ASV P90, in cation preference, pH optimum, and phosphatedonors(7). Similar to other ASV-transformed cells, organization of microfila-ment bundles in UR2-transformed CEF is significantly de-creased compared with that in uninfected cells (A. Antler, M. F.Greenberg, G. M.Edelman,and H.Hanafusa, person-al communication). An interesting question remains as to how UR2 inducesaunique,elongatedtransformed morphol-ogyinCEF.

*Correspondingauthor.

Clearly,theUR2rosrepresents amemberof the

family

of retroviraloncogenescoding forthetyrosine-specific protein kinases, despite differences in nucleotide sequences among

thesegenes. Tounderstand thebasis for the similarity and

difference

of thetransforming functionsamong rosand other

oncogenes, a detailed

analysis

of the

genetic

structure of UR2 androssequence is necessary.

We have molecularly cloned

the

full-length genomes of

UR2 and UR2AV. Both were shown to be

biologically

active. Cross hybridization among ros and three other ASV

transforminggenes (src of RSV, fps[22]of Fujinami sarco-ma virus, andyes [9] of Y73 ASV) showed that ros shared somesequencehomologywithfps and little or no homology

withsrc andyes.

MATERIALS AND METHODS

Cells and viruses. ThepreparationofCEF, UR2,and the

subgroup A UR2-associated virus (UR2AV) followed the previously

published

procedure (2, 31). A

methylcholan-threne-transformed quail cell line, QT6 (15), was cultured

similarly

to CEF.

Isolation of closed circular proviral DNA. QT6 cellswere seeded at 5 x 106 cells per dish and infected with

UR2(UR2AV)

at a multiplicity of infection of 1 in the

presence of 17 ,ug ofDEAE-dextran per ml. The medium

was replaced 5 hlater, and at 24 h postinfection the cells

were harvested and circularproviral DNA was isolated by

Hirt precipitation and

acid-phenol

extraction as described elsewhere(22, 29a). Ayieldof 960 ,ug of DNA was obtained

fromatotalof 98 8.5-cm dishes after acid-phenol extraction and was applied to aBio-Rad A5m (100- to 200-mesh) gel column to remove small, contaminating linear DNA

frag-ments. The column was washed, and the DNA was eluted with abuffercontaining 20 mM Tris-hydrochloride (pH 7.2), 0.1 N NaCl, and 5 mM EDTA. A total of20 ,ug of closed circularDNAwas recoveredin thepeak fractions.

914

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

Molecular cloning and subcloning of UR2 and UR2AV DNAs. One microgram each ofEcoRI- orSstI-cutUR2 and UR2AV closed circular proviralDNAs were ligated to two

micrograms of AgtWES *XB (11) EcoRI or SstI arms,

re-spectively. Onemicrogramof the ligation mix was packaged in vitro into lambda phage particles (11, 22). Lambda arms were purified from the internal fragment by sucrose gradient centrifugation (13). The recombinant phages were titrated on EscherichiacoliED8654 (16). Packaging efficiencies of 2.6 x

105

and 8 x 103PFU/,ug of DNA were obtained for the

SstI-and EcoRI-cut and -ligated DNAs, respectively. Screening of the recombinant phages with

32P-labeled

cDNA made

from 24S UR2 RNA was done by the procedure of Benton

and Davis (3).

Subcloning of the EcoRI insert from the UR2-lambda

recombinant phage clone into pBR322 was done

by

the

method of Bolivar et al. (4). The UR2

plasmid

clone was

called pUR2. Similarly, a750-base-pair(bp) EcoRI-PvuII v-rosspecific DNA fragment wascloned

by

using

the

2,293-bp

PvuII-EcoRI fragment of pBR322 DNA. This clone was called prosl. E. coli C600 cells

(1)

were used for the

transformation.

Transfection of cloned UR2 and UR2AV DNAs. UR2 and

UR2AV DNAs werefreed fromthe vectorDNAs

by

restric-tion enzyme digestion and were

purified

by

agarose

gel

electrophoresis. They weretransfected without

prior

ligation

onto CEF by the calcium

phosphate

method

(8)

with the

followingmodifications. UR2 and UR2AVDNA

(1

p.g each)

were mixed with salmon spermDNAtoatotalof 20 ,ug and

added to 500

RI

of 20 mM HEPES

(N-2-hydroxyethylpipera-zine-N'-2-ethanesulfonic

acid;

pH 7.0-137 mM

NaCl-5

mM

KCl-0.59 mMNa2HPO4. A 0.1 volumeof 1.25M

CaCl2

was added, and the mixture wasallowedto sitat room tempera-turefor 15 min. Thesolution wasthen addedtocells in 5 ml

ofgrowth medium, and the culturewas incubated at

37°C.

The cells were seeded 1 daybeforeat7 x

105

cellsper 6-cm plate.Themediumwaschanged5 haftertheadditionofviral DNAs. The cells weretransferred to 8.5-cm dishes as

they

became confluent, and they were overlaid with soft agar medium the nextday toenhance the

growth

of transformed

cells. Virus production was

assayed

by

determining

the

presence of reversetranscriptase

activity

inthesupernatants oftransfected cells (29).

Protein analysis. UR2/UR2AV-transfected cells grown in

6-cm dishes were labeled with 500

,uCi

of

[35S]methionine

perplate for 5 h. Extraction of cold or

35S-labeled

cellular

proteins,

immunoprecipitation,

protein

kinase assay, and

sodiumdodecylsulfate

(SDS)-polyacrylamide gel (5

to

15%)

electrophoresis followed previously described methods

(6).

Anti-gag serum and serum from an RSV-infected tumor bearing rabbit (TBR) were kindly

provided

by

Ricardo Feldman.

RNA blotting andhybridization.Totalcellular

polyadenyl-ic acid-containing RNAs from UR2AV- and

UR2/UR2AV

DNA-transfected cellswereisolated

by

previously

described

procedures (27, 30). RNAs were denatured with

glyoxal,

separated on 1% agarose gels, transferred to nitrocellulose

paper (27, 30), and hybridized to a5'

probe

containing

the leader sequence(seebelow)and toaprobederivedfromthe 750-bp ros-specific DNA described above. Conditions for hybridization andwashing ofthe filters have beendescribed

previously (30).

Preparation of

32P-labeled

DNAprobes. 24S UR2 and 35S

UR2AV RNAswere isolated from

purified

virus asdescribed

previously (28, 29),

polyadenylic

acid selected

twice,

and usedseparately astemplatesin the in vitroreverse

transcrip-tase

reaction.

The

200-pld

reaction mixture contained 50mM

Tris-hydrochloride (pH 8.0);

40 mM

KCl;

4 mM

MgCl2;

2

mM

dithiothreitol;

200

,uM

eachof

dATP,

dGTP,

and

TTP;

60,ugof calfthymusDNA

primers (26),

20,ugof

actinomycin

D, 300

,Ci

(3,000

Ci/mmol)

of

[32P]dCTP,

150 U of avian myeloblastosis virus reverse

transcriptase

(provided by

J.

Beard, Life

Sciences,

St.

Petersburg,

Fla.,

through

the

courtesyof J.

Gruber,

Resource

Program,

National Cancer

Institute),

and0.3 ,ugof

template.

Thereaction mixturewas

incubatedat

37°C

for2h, and the reactionwas

quenched

by

theaddition ofasolution

containing

0.5M

LiCl,

10 mM Tris-hydrochloride (pH 7.2), 10 mM

EDTA,

and

0.2%

SDS. After ethanol

precipitation,

the

sample

was base

hydrolyzed

with

0.2 N NaOH

containing

1 mM EDTA at

100°C

for 1

h,

neutralized,

and

passed

through

a

Sephadex

G-50 column.

TheDNAfrom the void-volume fractionswas

precipitated

in ethanol.

Preparation of the

following

probes

specific

to various

regions

of thegenomeof the

Schmidt-Ruppin

strain of RSV

was as described

previously (27).

The 5'

probe

isa

500-bp-long

fragment

extending

fromtheEcoRI site within the U3

region of the left-hand

long

terminal repeat

(LTR)

to the BamHIsite in the 5' gag

region;

the 5' gag

probe

is 1.3 kb and spans the two BamHI sites within the gag

region;

and

the 3'gag

probe

extendsfrom the secondBamHI site ingag to the downstream

EcoRI

site and is 400

bp

in

length.

The

pol and

pol-env probes

were

prepared

from

pSR2

(5).

The

polprobe is 1.45 kb and extends from the HindIII to the

BglII

site inpol; the

pol-env

probe

is 1.8

kb,

extends from

the

BglII

site inpoltotheEcoRIsite inenv, and coversthe

3'

portion

ofpol

plus

more than 1 kb ofthe env sequence.

The c

probe

covers the 450

bp

extending

fromthePvuII to

the

EcoRI

site in the U3

region

of

Schmidt-Ruppin

B.

A

probe

specific

to3'

v-fps

wasderivedfrom

pBRFO4,

a

plasmid

containing

a

400-bp

BamHI DNA

fragment

from

Fujinami

sarcoma

virusfps

(22).

The3'v-yes

probe

usedwas

the1.1-kb

PstI

fragment

from Y73

(9),

andthe 3'v-src

probe

used was a

900-bp

PvuII

fragment

from

pTT107

(25).

Two

ros-specific

probes

were used: one was the

750-bp

EcoRI-PvuII

fragment

whichcoversthe 3'two-thirds ofros,andthe

other was an internal ros sequence, the

300-bp

AvaI

frag-mentD

(see

Fig. 6).

Restriction

mapping.

Two

approaches

were used. Inthe first

approach,

end-labeled DNAs were used for

mapping.

UR2DNAisolated from

EcoRI-cleaved

pUR2

waslabeledat

its 5' end with

32P

by

using

T4

polynucleotide

kinase

(Bethesda

Research

Laboratories)

after

digestion

with

bacte-rialalkaline

phosphatase

(Bethesda

Research

Laboratories).

UR2AV DNA from one of the recombinant clones was isolatedfrom its lambdavector

by

digestion

with SstI. The3'

protruding

SstI endswerelabeled with 3'-dATP

([32P]cordy-cepin

5'

triphosphate;

New

England

Nuclear

Corp.)

and

terminal transferase

according

totheconditions

provided

by

themanufacturer. The end-labeled DNAswere

digested

with

a

single-cut

enzyme to generate two

fragments

of

unequal

size,

each

containing

a

single

32P-labeled

end. These

frag-ments

(104

cpm

each)

were

subjected

to various restriction enzyme

digestions,

and

portions

of the reaction were re-moved 3, 6, 10, and 60 min afterinitiation of the

reaction,

which was

stopped

by

the addition of EDTA to 10 mM followed

by

heating

at

65°C

for 3 min. About

5,000

cpm of

each

restricted,

end-labeled

fragment

was

electrophoresed

through 5%

acrylamide

and

1.4%

agarose

gels,

using

32p_

labeled

HindIII-digested

lambda DNA and

Hinfi-digested

pBR322

DNA as markers. The

gels

were dried on DE81

paper

(Whatman,

Inc.)

and

exposed

for 8 to 12 h with a

on November 10, 2019 by guest

http://jvi.asm.org/

(3)

916 NECKAMEYER AND WANG

.I

a

m 0 C.) (I)

kb

9.6-

6.7

-kb c

B

.~z

> 0

z < <

0 c'. C'.

> a: c 4: .-.. 1-1

('C4 (\4 a: a: a

D D

9.6- P* _{_}

_{_F4}

_

(: <

.1- z

>;0

('4 ('4

a: cc

co N

C: a: D

0 -35S

4.3-2.2- *

1.9-U.

FIG. 1. Restriction enzyme analysis of circular UR2 and UR2AV DNAs. Material containing unintegrated viral DNA (5 ,ug) was restricted with either SstI or EcoRI, and each digest (2 ,ug) was loaded per well, electrophoresed through a 0.8% agarose gel, blotted ontonitrocellulose paper, and probed with32P-labeledcDNA made from 24S UR2 RNA.

Cronex intensifying screen (Du Pont Co.). In this manner, theorder ofrestriction sites from the labeled end could be

precisely mapped.

In the second approach, cold UR2 and UR2AV DNAs weredigested completely with the same set of enzymes used

for partial mapping, electrophoresed through 0.8% agarose

gels, and blotted onto nitrocellulose (23). The approximate gene boundaries were determined by hybridizing the blots

with 2 x

105

to 5 x

105

cpm of the gene-specific probes

mentioned above. Hybridization and washing conditions

havebeen described elsewhere (21).

Hybridization of v-ros to 3'-specific v-onc probes. UR2

DNAwasdigested withenzymesthatcutwithinros,

gener-ating 5'- and 3'-ros

specific

fragments, which were blotted and hybridized to 2 x

106

cpm of probe under low

(35%

formamide,

5x SSC [lx SSC, 0.15 M NaCl

plus

0.015 M

sodium citrate]) or moderate

(50% formamide,

3x

SSC)

stringency at 37°C for 2 days.

Washing

conditions for moderate-stringency hybridizations were similar to those described before (21); for

low-stringency

conditions,

blots

werewashed three times for20min eachin 300 ml of 26mM

Tris-hydrochloride

(pH 7.4}-2x SSC-1 mM EDTA-0.1%

SDS at 55°C, followed by a similar wash at 60°C. For sequential hybridization, the previous probe was eluted at

68°C in 40 ml of 50% formamide-1x SSC-50 mM Tris-hydrochloride (pH

7.4)-i

mM EDTA-0.1% SDS. The blot was exposed to an X-ray film overnight to determine the extentof elution.

RESULTS

Restriction enzyme analysis of circula

UR2

and UR2AV DNAs. Unintegated viralDNAisolated from

UR2(UR2AV)-infected QT6 cells was subjected to digestion by various

restriction enzymes to determine which endonucleases cleaved at a single site and thus would be suitable for

cloning. The restriction patterns of EcoRI and SstI diges-tionsare shown inFig. 1.

SstI

cutUR2 and UR2AV DNAs

2.2

t1.9-40

-24S

-22S

5'

probe

ros-specific

probe FIG. 2. RNAs of UR2 and UR2AV DNA-transfected CEF. Polyadenylic acid-containing RNAs (10 ,ug) isolated by SDS-pro-teinase K extraction of UR2- and UR2AV-transfected cells were denatured with 1 Mglyoxal andfractionated on 1% agarose gels. TheRNAs weretransferred to nitrocellulose paper and hybridized with a 5' leader or ros-specific probe. RNAs isolated from UR2(UR2AV)-infected cells were used as controls. 32P-labeled HindIII-cut lambda DNAs were denatured and run in parallelfor molecularweight markers.

each only once, and the linearized viral DNAs containing either one or two copies of the LTR banded at

positions

corresponding

toca.

3.5

kbfor UR2 and 8.0 kbfor UR2AV.

EcoRI cleaved UR2 DNA once, liberating the linearized UR2 DNA, but cut UR2AV DNA three

times, releasing

several subgenomic fragments. The majority of the acid-phenol- and column-purified viral DNAs appeared to be of the circular form. However, the majority of the DNA

preparation was of cellular origin, because the viral DNA

could be detectedonly

by hybridization

tothe cDNA

probe

andnot by ethidium bromide staining ofthe 2

jig

of DNA loadedperwell in the agarosegel(Fig. 1).

Molecularcloning of UR2 and UR2AV DNAs. The

remain-ing

DNA from theEcoRI andSstI

digests

described above was used for cloning, using purified XgtWES - XB EcoRI

and SstIarms. cDNAprobe made from24SUR2 RNAwas used for the screening. Eight full-length helper viral DNA clones were obtained, using the SstI site for the

cloning.

Eleven lambda-UR2AV recombinantphages, each contain-ing more than one UR2AVEcoRI fragment,and a

lambda-UR2 clonewere

isolated,

usingthe EcoRI site for

cloning.

About

0.15%

(12 clones of8,000 plaques screened) ofthe

recombinantphages containedviralsequences,

using

EcoRI arms ofthelambdaDNAforthe

cloning.

The number of UR2AV clones isolated was

roughly

10-foldthe numberof UR2 clonesisolated.Thiscanbe expect-ed from the ratio ofUR2AV to UR2 DNA

(Fig.

1). This

probablyreflected theratio of thehelpertothe UR2 virus in the stock used forthe infection of the QT6cells. This has

been

previously

observed in several UR2(UR2AV) virus

mikowumo. .K

'59IFFiRlw-.gz

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.612.336.550.72.330.2] [image:3.612.136.231.73.281.2]

(4)

MOLECULAR CLONING OF AVIAN SARCOMA VIRUS UR2 917

UR2/UR2AV

control

virus DNA

I

l1---o

TBR

gag

TBR

gag

TBR

gag

-Pr180

e

_

cmw eini clw

t_

~~p

7₇₆

2l

_p6

'uur_up27

dUl

E_

_S

CU-p19

[image:4.612.64.286.72.397.2]

cm' ClEW

cici-p15

pl2

FIG. 3. 35S labeling ofproteins from UR2 and UR2AV DNA-transfected cells. Cell lysates prepared from UR2- and UR2AV-transfected CEF that had been labeled for 5 h with 500 ,uCi of

[35S]methionine

per 6-cm plate were immmunoprecipitated with

TBR oranti-gagserum.The immunecomplexeswere subjectedto electrophoresis on an SDS-5 to 15%polyacrylamide gradient gel.

3"S-labeled

proteins from cell lysates of uninfected and

UR2(UR2AV)-infected CEFwereanalyzed inparallel.

stocks (2). Each UR2 and UR2AV lambda recombinant phage clone was purified by three to four cycles of single

plaque isolation.

Toconfirm the identity of the UR2 clone, the viralDNA

insertwashybridized with probes specifictovariousregions

of the RSV genomeby the procedure of Southern (23). As

expected, the UR2 DNA hybridized onlyto the 5', 5' gag,

and c probes, in addition to the cDNA made from UR2

genomic RNA,butnottothepoland pol-env probes. It has

beenpreviously determined(31) thatthe pol geneisdeleted and the gag and env genesaretruncated in the UR2 genome. Because the size ofthe insertfromthis recombinant clone,

3.4kb, isequivalenttothat determined for the UR2 genomic RNA (31), it is most likely that this clone contains a

full-length copy of the UR2 genome. However, this clone contained, in addition to the 3.4-kb UR2 insert, atandemly

linked 3.7-kb DNA fragment of nonviral sequence.

There-fore,the3.4-kb UR2 DNAwaspurifiedand subcloned into the EcoRI site ofpBR322. All further studies of the UR2 genome were done with this recombinant plasmid clone, called pUR2.

Transfection assays of UR2 and UR2AV DNAs. Insert DNAs isolated from pUR2 and from one of the UR2AV

lambda clones, 4B2, were transfected together onto CEF.

UR2/UR2AV control

gag TBR

virus

gag TBR

'._

UR2/UR2AV DNA

Ir

gag TBR

-p68

_-IgG

(53K)

FIG. 4. Protein kinase assay of UR2 and UR2AV DNA-trans-fected cell extracts.Celllysates of CEF transfected with UR2 and UR2AV DNAs were immunoprecipitated with TBR or anti-gag serum andincubated in vitro with

[_y-32P]ATP.

The phosphorylated proteins were analyzed on an SDS-8.5% polyacrylamide gel. Cell lysates prepared from uninfected and UR2(UR2AV)-infected CEF were treated similarly and used ascontrols. IgG, Immunoglobulin G.

The transfected cells displayed

typical

UR2 transformed elongatedmorphology andproduced UR2(UR2AV)

pseudo-type ca. 2weeks after transfection. This demonstrated that

both UR2 and UR2AV clones tested were

biologically

active. In addition, five independently isolated UR2AV-lambda clones(A21,

Dl,

1C-1, 4-1, 14-1)were cutwithSstI tofreetheinsert from thelambdaDNAsandwere

transfect-ed similarly onto CEF. Assays for reverse transcriptase activity in the culture fluid of transfected cells 10

days

posttransfection

were

positive

for all clonesexceptclone

1C-1. The map of the UR2AV DNA insert of clone

iC-1

was

identical to that of 14-1, which was

biologically

active,

exceptthat the

iC-1

insertwasmissinga

portion

of the SstI-HindIl

right-hand

region

(see

Fig. 5).

This

region

contains

theLTR, anddeletion ofthis segmentmightaccountforthe

loss of biological activity.

Toconfirmthat thetransformation wasinducedby UR2, totalpolyadenylic acid-containing cellularRNAwasisolated

from thetransfected cells andanalyzed byRNAblottingand hybridization. The transfected cells displayedasubgenomic

andgenomicRNA patternidenticaltothatof

UR2(UR2AV)-infectedCEF(Fig. 2). Hybridization ofthe RNAs tothe 5'

probe detected 35S genomic RNA and 22S subgenomic envelope mRNA of UR2AV, as well as the 24S

genomic

UR2 RNA. Cells transfected with UR2AV alone produced only 35S and 22S RNAs. A

ros-specific probe hybridized

only with the 24S UR2 RNA. The RNA patterns of all

UR2AV-transfected cells were

identical;

however, as ex-VOL.50, 1984

.4n

.i;

"M

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.612.321.551.73.361.2]

(5)

pected from thereverse transcriptaseassay, no viralRNAs were detected in clone 1C-1-transfected CEF.

Viral specific proteins were also analyzed. CEF

trans-formed by the molecular clones were labeled with

[35S]methionine,

and cell lysates were precipitated with

RSV-infected TBR oranti-gag serum (Fig. 3).

DNA-trans-fected andvirus-infected cellsgavesimilar patterns with the

same antisera, which precipitated the gag-rosfusion

prod-uct, P68.Recognition ofP68byTBR serum wasapparently via the gag peptide in P68. The appearance ofP68 as a

doublet in virus-infected cells has been observedpreviously (7). The reasons for theappearance of this doublet arenot

clear. For assaying the kinase activity, cold cell lysates prepared fromUR2/UR2AV-transfected CEFwere immuno-precipitated with TBR or anti-gag serum and incubated in

vitro with

[y-32P]ATP.

The results of this assay indicate that P68 was associated with protein kinaseactivity (Fig. 4). The lower level of P68 kinase activity in the transfected celllanes ascompared with that in virus-infected cells wasdueto the

fact that only ca. 20% of the transfected cells were

trans-formed at thetime of the experiment, as opposed to

com-plete transformation in the virus-infected culture. The lower

degree oftransformationcould be due tointerference of the

helperviruswith the spreadingof UR2(UR2AV). The non-P68 bands seen in the virus-infected culture (Fig. 4) were most likely due to incomplete washing of the immunocom-plex. It was observed previously that immunoglobulin G could be phosphorylated by P68, particularly when TBR serum wasused in the immunoprecipitation (7).

Theabove studies show that molecularly cloned UR2 and UR2AV DNAsare biologically active andindistinguishable in effectonCEFfrom theirrespective parental viruses.

Restriction maps of UR2 and UR2AV DNAs. Restriction

maps of UR2 and UR2AV were constructedby two meth-ods.Inthefirstapproach,

32P-end-labeled

DNA was

partial-ly digested with restriction enzymes and analyzed by gel electrophoresis.Inthe secondapproach, restriction

enzyme-digested, unlabeled DNA washybridized withgene-specific probesto determine the physical order and genetic content

of the DNA fragments. The results of such mapping for

UR2AVand UR2DNAs areshown in

Fig.

5and6,

respec-tively.

The UR2AV DNA containing two LTRs is 8.0 kb, and several of its restriction sites are similar to those of

Rous-associated virus-2(RAV-2)(17). The KpnI, SstI, andBamHI

sites in RAV-2are conserved in UR2AV. AHindIII site in

RAV-2 is also conserved in UR2AV, although UR2AV containstwo extraHindIIIsitesin theenv

region.

Similarly,

S B

UR2AV ' B E.1 H X B

90g pO/

EK B EH H

'II *UW-'-FM env

S B H S

UR2

'i

n

agog ros Aenm

0 2 4 6

UR2AV shares with RAV-2 two EcoRI sites, yet has an

additionalEcoRIsite inthemiddle ofthe UR2AVgenome. It

doesnotcontainanEcoRI site intheLTR, as does RAV-2. By comparing the restriction maps of UR2AV and UR2

DNAs, thegag-rosandros-envbordersmaybedetermined. Construction ofa detailed map ofUR2 (Fig. 6) allowed a more precise definition of the gene borders than those

determinedbefore (31). UR2 DNAhas asingle EcoRI site.

There is no corresponding EcoRI site in UR2AV DNA,

suggesting that this site is located within the ros-specific

sequence. The gag-rosboundary appears tobe locatedvery

closetothe HaeIIsite immediatelyupstreamfrom the

right-hand EcoRI site, because the 460-bp HaeII fragment B

hybridized strongly to a UR2 representative probe and weakly to a UR2AV representative probe and the 5' gag

probe (data not shown). The 750-bp PvuII fragment B

appearedtocontainonly therossequencesince it hybridized onlytothe UR2, butnottheUR2AV, cDNAprobe.The ros-env boundary was mapped between the PvuII site and the

second AvaI site from the left, because AvaI fragment B

hybridizedtoboththe UR2 and UR2AV cDNA probes (data notshown). Since AvaI fragmentBcould not hybridize with gag,pol, andcprobes, weinferredthat the sequence present

in this DNA and hybridizable with UR2AV cDNA must be

the env sequence. The mapping is consistent with the

previous finding that two highly conserved env-specific

oligonucleotides, spots 11 and 12a, located at 96 and 594

nucleotides, respectively, upstream from the termination

codon of gp37 of the SR-Agenome, were presentin the UR2

genomicRNA(31).Given the estimatesof these borders, ros

isca. 1.2kb in length.

From the results ofthe restriction enzyme analysis, we

concluded that the UR2 genome is 3.4 kb in

length

and contains inthe middle ca. 1.2 kb oftransforming sequence ros.UR2 shares with UR2AV 0.8kb of5'leaderand gag and 1.4kbof3' env andcsequences.

Homology between v-ros and other transforming genes.

of therossequence with those ofsrc,

fps, andyesby

liquid

hybridization

between viral RNAsand cDNAs

specific

to individual v-onc sequences detected no

significant

homology between ros and the rest ofthe ASV

Eco RI U3RU5 EcoF

L2 ,,-n

-ros v4e I

SonHI

8gl/

S Hoe31

Di MIncl HuindE NciI Nco I NruI

- Kb SmoI

Ssf

I

FIG. 5. Restriction map of cloned UR2AV DNA. The 8.0-kb UR2AV proviral DNAwas purifiedfromUR2AV clone 14-1, and restrictionenzyme siteswere determined. S, SstI; B, BamHI; E, EcoRI;X,XbaI; K, KpnI; H, HindIll. Thegenetic structure was determinedbyhybridizationof the UR2AV restrictionfragmentsto

probesspecific for varioussequencesof the RSVgenome.TheUR2 DNA was included here forcomparison to show the helper virus-related androssequences.

S/u]

D3001

B550 A1810

If|

C460 |E

A2690 8710

B A3275

A2690

AC2501

B460

8480

|A2860

B1425

I

A1975

C490 A2150 D| B700

B1550 A1850

81300 A2100

8750 A2650

A2700 B700

A2450 B950

B380 C380 A2640

FIG. 6. Restrictionmapof cloned UR2 DNA. The 3.4-kb UR2 DNA insert was isolated from pUR2, and restriction sites were determined as described in the text. The numbers in the boxes indicatethelength inbp oftheindividual restrictionfragments.The sizes ofrestrictionfragmentsAvaI-E, AvaI-F, BglII-B,andNciI-D are160,120, 125,and 60bp, respectively.

J. VIROL.

RI

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.612.327.558.509.673.2] [image:5.612.66.308.576.655.2]

(6)

MOLECULAR CLONING OF AVIAN SARCOMA VIRUS UR2 919

a. Z > oa:

A

5 E

a

I IL z > omID

B

=) m n

, Gw

UJ

a) 0

0r a.

, L Z > oE X

C D n L

< - J¶

V

a

m

a

35%formamide 5XSSC

v-yes specific 35%v-srcformamidespecific5XSSC

50% formamide 3X SSC

v-ros specific

_>

>CD

0r a. < m

IL

X~a: a: a:

at

-750

-300 -125

50%formamide3X SSC

v-fps specific

[image:6.612.130.482.73.473.2]

50%formamide 3X SSC v-ros specific

FIG. 7. Homology betweenv-rosand 3'v-oncregions. UR2 DNA insertwasrestricted withenzymescleavingwithinrostoyield5'-and

3'-rosspecific fragments. pTT107andpsrc6 src-containing plasmids (see text)and thepBRFO4fps-containing plasmidDNAwerecutwiththe appropriateenzymetofree theoncinsert. Theamountof eachDNAaddedperwellwasadjustedtoyieldca.50ngof eachoncDNAfragment containingtheputative regionofhomology. The DNAswerefractionatedon 1%agarosegelsandtransferredtonitrocellulose filters. The probesused ineachhybridizationwere:(A)1.1-kb PstIfragmentfrom lambda-Y73clone, (B)0.9-kb PvuIIfragment frompTT107, (C)0.75-kb

PuvII-EcoRI fragmentfrompUR2, (D)0.4-kb BamIinsertfrompBRFO4,and(E)0.3-kb AvaI-EcoRIfragmentfrompUR2.

transforming genes (21). However, it is possible that the homologymaybe restrictedtoasmallregion ofrosand thus

cannot be easily detected by hybridization between total viral RNA and the cDNA probe representing the entire domain ofav-onc gene. To test thispossibility, we under-took the approach of subdividing the ros sequence by

digestionwithenzymes knownto cutwithinrosand

hybrid-izing the individual fragments with probes prepared from other v-onc DNAs. Probes derived from the 3' regions of

src,fps, andyes were used since theseregionswereshown

tocontain thesequencesconservedamongthose transform-inggenes(9,20, 22) (Fig. 7). No significant hybridizationwas

detected betweenv-rosandv-yes, evenunder conditions of

low stringency (35%formamide, 5x SSC), although the v-yesprobe cross hybridized with the 3.1-kb EcoRI fragment

ofSR-A(frompTT107) containingtheentirev-srcsequence, aswellaswith the0.75-kb PstI 3'-srcspecific fragment (from

psrc6). The high degree of amino acid sequence homology

within the C-terminal half ofpp60rc and pp9OYes has been shown previously (9). Hybridization of this v-yes probe to UR2 DNAfragments containinggagandenvsequencesand

to UR2AV DNA is apparent in Fig. 7A. This was due to contamination of the yes-specific probe with helper virus-relatedsequencesofY73, since the 1.1-kb 3'v-yesfragment was slightly contaminated with the 1.4-kb gag-5' yes

frag-ment and the 1.5-kb fragment containing the env and LTR

region of Y73 (9). The 1,300-bp NruI fragment B (Fig. 6) hybridized to this 3' yes probe, apparently due to the env sequences presentin this DNA. Neither of the ros-specific fragments (HincII-B and PvuII-B) hybridized with the 3'-yes probe.

Nohybridizationbetweenv-srcandv-rossequencescould be detected, although intense hybridization of the v-src

probe to pTT107EcoRIand psrc6 PstIwas seen (Fig. 7B).

Theupperband of the doublet in thepsrc6lanerepresented the partially digested, linearized psrc6; the lower band,

I

-8000

41

-1300

-750

-530

L I

mE CL

D

X

a

.0 75

N _

a: :c

bp

I

VOL. S0, 1984

on November 10, 2019 by guest

http://jvi.asm.org/

(7)

which hybridized with much less intensity, was the pBR322 vector. The insert src DNA used as probe might not have been completely purified from the pBR322 vector after one cycle of gel purification. The

EcoRI-PvuII

ros-specific probe hybridized to the expected ros-containing, but not to the src-containing, DNAs (Fig.

7C).

However, under low as well as under moderate (50% formamide, 3x SSC) stringency, the 400-bp probe repre-sentative of the 3' conserved region of v-fps (the BamHI insert from

pBRFO4)

(22) hybridized significantly to v-ros (Fig. 7D).

AvaI

cleaves UR2 into six fragments (Fig. 6), four of which contain portions of the entire v-ros: the 460-bp

AvaI-C

covers the gag-ros junction, including ca. 300 bp of

5' v-ros; the 160-bp AvaI-E and the 300-bp AvaI-D contain only internal v-ros sequences; and the 550-bp AvaI-B con-tains ca. 450 bp of 3' ros plus env sequences. Only the 300-bp

AvaI-D

hybridized with the v-fps probe. Conversely, a probe

made from AvaI-D hybridized to the 400-bp 3' v-fps DNA fragment (Fig. 7D and E).

BglII

cleaves UR2 DNA into the 125-bp B fragment that covers the 5' portion ofAvaI-D and into the 3,275-bp A fragment containing the rest of the genome (Fig. 6). The v-fps probe hybridized only to the 3,275-bp BgIII fragment A (Fig. 7D), indicating that the sequences homologous to v-fps were contained within a

175-bp ros sequence between theBglIIand AvaI sites. However, we cannot rule out the possibility that failure to detect BglII fragment B might be due to the inefficiency of DNA transfer. The location of the fps-related sequence within v-ros is

reminiscent of that of v-abl (18) among the transforming genes coding for tyrosine protein kinases. We checked the homology between v-ros and v-abl. When a probe derived from the 5' 1.2-kb of v-abl DNA (19) was used, no significant hybridization could be detected (data not shown).

DISCUSSION

Biological activity of the UR2 and UR2AV clones. Although UR2 DNA could readily induce transformation of CEFwhen cotransfected with UR2AV DNA, so far we have not been able to transforma rat cell line,3Y1, with UR2 DNA. Under similar conditions,3Y1 cells could be transformed byeither

SR-A or Fujinami sarcoma virus DNA (14, 22;unpublished data). The reason for the failure of UR2 DNA to transform 3Y1 rat cells is not clear. We are currently investigating the possibilities that the promoter sequence in UR2 DNAmight

not be recognized efficiently in the rat cells orthat a second oncogene may be required for ros to induce complete transformation of rat cells (10).

Seven of the eight UR2AV clones were tested for biologi-cal activity by transfection onto CEF. All but one clone, UR2AV iC-1, were found to be biologically active. Clone UR2AV

1C-1

was shown by preliminary restriction mapping to be truncated in the region of the LTR. This deletion was most likely responsible for the loss ofbiological activity.

Genetic structure and gene product ofUR2. Hybridization of the UR2 DNA restriction fragments toprobes represent-ing different regions of the SR-A genome and sizing of the fragments have enabled us to map more precisely the genomic domains of UR2 than thoseobtainedpreviously by oligonucleotide mapping (31). Ourdataindicate that the UR2

genome contains ca. 0.8kb of 5' leader and gagsequences,

1.2 kb of ros-specific sequence, and 1.4 kb of env and 3' sequences. UR2 has been shown tocodefor agag-rosfusion

protein of 68,000 daltons (P68)withtyrosine-specific protein

kinase activity (7). Our data suggest thatthe partialgag and the ros sequences code for 15- and 44-kilodalton

polypep-tides, respectively.

Thisleavesca. 9kilodaltons of

peptides

unaccounted for in P68. Itis possible that the residualenv

sequences may code for the 9-kilodalton of the C-terminal

peptides

of P68. Alternatively, the apparent molecular

weight of P68 determined by SDS-polyacrylamide gel

elec-trophoresis

may notreflect thetruevalue dueto

glycosyla-tion orunusual amino acid sequence. Glycosylation of the

rosfusion proteinseemsunlikely sinceweshowed

previous-ly

that in vitro translation of the 24S UR2 genomic RNA,

using

reticulocyte lysate, only yieldedthe P68 gag-rosfusion

protein

(7).

Homology ofros with other oncogenes. Our data indicate thata sequence within the 300-bp AvaIfragment D (Fig. 6) hassignificanthomology with the 3' conserved regionoffps. This is unusual among the tyrosine protein

kinase-coding

transforminggenesinthatthe conserved region in v-ros is at

the middle of the gene. Theonlyother retroviral

transform-ing

gene of the tyrosine kinase family with conserved

sequenceoutside the 3' region is v-abl, and v-ros apparently

shares no homology with v-abl. It has been shown thatthe protein kinase catalytic domains of pp60src,

pp140fPs,

and

ppOWYes

are contained within the carboxy-terminal half of these proteins (12, 20, 32; J. Brugge and D. Darrow, J. Biol.

Chem.,

in press). It is

likely,

given the middle conserved

region in ros, that the catalytic domain is contained within this domain of P68. Deletion of the 3' two-thirds of v-abldid

notaffect itstransforming ability (18). If the catalytic domain

ofv-ros isindeed located in the 5' half, itwill beinteresting

todetermine whether the 3' ros sequence is required forthe transforming function.

ACKNOWLEDGMENT

WethankHidesaburo Hanafusa and MariusSudol for the critical reading of the manuscript, David Foster for the gift of X-yes and pBRFO4 fps DNAs, Shinji lijima for the gift of psrc6 DNA, and Steve Goff for theablDNA.

Thisworkwassupported byPublic Health Service grant CA29339 to L.-H.W. from the National Cancer Institute. W.S.N. was sup-ported by Public Health Service training grant T32AI07233. L.-H.W. isarecipient ofa Public HealthService Research Career Development Award CA00574fronmthe National Cancer Institute.

LITERATURE CITED

1. Appleyard, R. K. 1954. Segregation of new lysogenic types dutinggrowth ofadoubly lysogenic strain derived from Esche-richia coli K12. Genetics 39:440-552.

2. Balduzzi, P.C., M. F. D. Notter, H. R. Morgan, and M.Shibuya. 1981. Some biological properties of two new avian sarcoma viruses. J. Virol. 40:268-275.

3. Benton, W. D., and R. W. Davis. 1977. Screening Xgt recombi-nant clones by hybridization to single plaques in situ. Science 196:180-182.

4. Bolivar,F. R.,L.Rodriguez,P.J. Greene, M.C.Betlach, H.L. Heyneker, and H. W. Boyer. 1977. Construction and character-ization of new cloning vehicles. II. A multipurpose cloning system.'Gene 2:95-113.

5. DeLorbe, W. J., P. A. Luciw, H. M. Goodman, H. E. Varmus, andJ.M. Bishop. 1980. Molecularcloning andcharacterization of avian sarcomavirus circular DNA molecules. J. Virol. 36:50-61.

6. Feldman, R. A., T. Hanafusa, and H. Hanafusa. 1980. Charac-terization of protein kinase activity associated with the trans-forming gene product ofFujinami sarcoma virus. Cell 22:757-765.

7. Feldman, R. A., L.-H. Wang, H. Hanafusa, and P. C.Balduzzi. 1982. Avian sarcoma virus UR2 encodesatransformingprotein which is associated with a unique protein kinase activity. J. Virol. 42:228-236.

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

(8)

8. Graham,F.L., andA.J.vander Eb.1973. Anewtechnique for the assayof infectivityof human adenovirus 5 DNA. Virology 52:456-467.

9. Kitamura, N., A. Kitamura, K.Toyoshima, Y. Hirayama, and M.Yoshida. 1982. Avian sarcoma virus Y73 genome sequence andstructural similarity of its transforming gene producttothat of Rous sarcoma virus. Nature(London) 297:205-208. 10. Land,H.,L. F.Parada,and R. A.Weinberg. 1983.Tumorigenic

conversion of primary embryo fibroblasts requiresatleasttwo cooperating oncogenes. Nature (London) 304:596-602. 11. Leder, P.,D.Tiemeier,and L.Enquist. 1977.EK2derivatives of

bacteriophagelambda useful in the cloning ofDNAfromhigher organisms: the AgtWES system. Science 196:175-177. 12. Levinson, A. D., S. A. Courtneidge, and J. M. Bishop. 1981.

Structural and functional domains of the Rous sarcomavirus transforming protein (p6Osrc). Proc. Natl. Acad. Sci. U.S.A. 78:1624-1628.

13. Maniatis, T.,R.C.Hardison,E.Lacy, J. Cauer, C.O'Connell, D.Quon, G. K. Sim,andA.Efstratiadis. 1978. The isolation of structural genes fromlibraries of eukaryotic DNA. Cell 15:687-701.

14. Mathey-Prevot, B.,H.Hanafusa,and S.Kawai. 1982.Acellular protein is immunologically cross-reactive with and functionally homologous to the Fujinami sarcoma virus transforming pro-tein.Cell 28:897-906.

15. Moscovici, C.,M.G.Moscovici,H.Jimenez,M.M.C.Lai,M.J. Hayman, and P. K. Vogt. 1977.Continuous tissue culture cell linesderived from chemically induced tumors of Japanesequail. Cell 11:95-103.

16. Murray, N. E.,W.J. Brammar,and K.Murray. 1977.Lamboid phages thatsimplify the recovery ofin vitrorecombinants.Mol. Gen. Genet. 150:53-61.

17. Payne, G. S.,S.A.Courtneidge,L.B.Crittenden,A.M.Fadly, J. M. Bishop, and H. E. Varmus. 1981. Analysis of avian leukosis virus DNA and RNA in bursal tumors: viral gene expression is notrequiredfor maintenance of thetumorstate.

Cell 23:311-322.

18. Prywes, R., J. G. Foulkes, N. Rosenberg, and D. Baltimore. 1983. Sequences of the A-MuLV proteinneeded forfibroblast andlymphoid cell transformation. Cell 34:569-579.

19. Reddy,E.P.,M.J. Smith,and A. Srinivasan. 1983.Nucleotide sequenceofAbelson murine leukemia virus genome: structural similarity of its transforminggene tootheroncgene products with tyrosine-specific kinase activity. Proc. Natl. Acad. Sci. U.S.A. 80:3623-3627.

20. Shibuya, M., and H.Hanafusa. 1982. Nucleotide sequence of Fujinamisarcomavirus: evolutionary relationshipof its

trans-forminggene withtransforminggenesof othersarcomaviruses. Cell30:787-795.

21. Shibuya, M., H. Hanafusa, and P. C. Balduzzi. 1982. Cellular sequencesrelated to three new onc genes of avian sarcoma virus (fps, yes, and ros) and their expression in normal and trans-formed cells. J. Virol. 42:143-152.

22. Shibuya, M., L.-H. Wang, and H. Hanafusa. 1982. Molecular cloningof the Fujinami sarcoma virus genome and its compari-son with sequences of other related transforming viruses. J. Virol. 42:1007-1016.

23. Southern, E. M. 1975. Detection ofspecific sequences among DNAfragmentsseparated by gelelectrophoresis. J. Mol. Biol. 98:503-517.

24. Takeya,T.,andH. Hanafusa. 1983. Structure and sequence of thecellular gene homolog to the RSV src gene and the mecha-nism forgenerating the transforming virus. Cell 32:881-890. 25. Takeya, T., H. Hanafusa, R. P. Junghans, G. Ju, and A. M.

Skalka. 1981.Comparison between the viraltransforming gene (src) of recovered avian sarcoma virus and its cellular homolog. Mol.Cell Biol. 1:1024-1037.

26. Taylor, J. M., R. Illmensee, and J. Summers. 1976. Efficient transcription of RNA into DNA by avian sarcoma virus poly-merase. Biochim. Biophys. Acta 442:324-330.

27. Wang, L.-H., M. Beckson, S. M. Anderson, and H. Hanafusa. 1984. Identification of the viral sequence required for the generation of recovered avian sarcoma viruses and character-ization of a series of replication-defective recovered avian sarcomaviruses. J. Virol. 49:881-891.

28. Wang,L.-H., and P.Duesberg. 1974. Properties andlocation of poly(A) in Rous sarcoma virusRNA. J. Virol. 14:1515-1529. 29. Wang, L.-H., P. Duesberg, K. Beemon, and P. K. Vogt. 1975.

Mapping RNase T1-resistant oligonucleotides of avian tumor virus RNAs: sarcoma-specific oligonucleotides are near the poly(A) end and oligonucleotides common to sarcoma and transformation-defective viruses are at the poly(A) end. J. Virol. 16:1051-1070.

29a.Wang,L.-H., B.Edelstein, and B. J. Mayer. 1984.Induction of tumors and generation of recovered sarcoma viruses by, and mapping of deletions in, two molecularly cloned src deletion mutants. J.Virol. 50:904-913.

30. Wang, L.-H., R. Feldman, M. Shibuya, H. Hanafusa, M. F. D. Notter,andP. C.Balduzzi. 1981.Genetic structure, transform-ing sequence, and gene product of avian sarcoma virusURL.J. Virol.40:258-267.

31. Wang, L.-H., H.Hanafusa, M. F. D. Notter, and P. C.Balduzzi. 1982. Genetic structure and transforming sequence of avian sarcomavirus UR2. J. Virol.41:833-841.

32. Weinmaster, G., E. Hinze, and T. Pawson. 1983. Mapping of multiple phosphorylation sites within the structural and catalyt-ic domains of the Fujinami avian sarcoma virus transforming proteins. J. Virol. 46:29-41.