Number and location of mouse mammary tumor virus proviral DNA in mouse DNA of normal tissue and of mammary tumors.

(1)

0022-538X/80/03-1013/13$02.00/0

Number and Location of Mouse Mammary Tumor Virus

Proviral

DNA in

Mouse DNA of Normal Tissue and of

Mammary

Tumors

BERNDGRONERAND NANCY E. HYNES*

Swiss Institutefor ExperimentalCancerResearch, Lausanne,Switzerland

The Southern DNA filter transfer technique was used to characterize the

genomiclocation of themousemammary tumorproviral DNA in different inbred

strainsof mice. Twoof the strains (C3H and CBA) arosefrom a crossofaBagg

albino (BALB/c) mouseandaDBAmouse.Themousemammary tumor

virus-containing restriction enzyme DNAfragments ofthese strains had similar

pat-terns,suggesting that the proviruses of these miceareinsimilar genomic locations.

Conversely, the pattern arising from the DNA of the GR mouse, a strain

genetically unrelatedtotheothers, appeared different, suggesting that itsmouse

mammary tumorprovirusesarelocatedindifferent genomic sites. Thestructure

of another gene, thatcoding for

fl-globin,

wasalso compared. The micestrains

whichwestudiedcanbecategorizedintotwoclasses,expressing eitherone or two

,8-globin

proteins. The macroenvironment of the

fl-globin

geneappeared similar

amongthe mice strainsbelongingto onegenetic class. Female mice of the C3H

strainexogenously transmitmousemammary tumorvirus viathemilk, and their

offspring have a high incidence ofmammary tumor occurrence. DNA isolated

from individual mammary tumorstaken from C3Hmice or from BALB/c mice

foster nursedonC3H motherswasanalyzed by theDNAfilter transfertechnique.

Additionalmousemammary tumorvirus-containing fragmentswere found in the

DNAisolated from eachmammary tumor. These proviral sequences were

inte-gratedintodifferent genomic sitesin each tumor.

Mouse mammary tumor virus(MMTV) is an

RNA-containing virus which causes mammary

tumorsinmice.Incertaininbred strains of mice

the virus isexogenouslytransmitted tothe

off-spring via the milk. Animals which receive this

milk-bornevirus haveahigh incidence of

mam-mary tumor occurrence.Inaddition, mice

trans-mit MMTV-specific information to their

off-spring via the germ cell. This endogenously

transmittedMMTVispresent asproviralDNA

integrated into the genomicDNAofallcells of

themouse (2,23,27).

All tested inbred strains of mice contain

en-dogenously transmitted MMTV proviral

se-quences (22, 25, 35, 44). The origin and the

possiblefunction of proviralDNA remain

spec-ulative. Inallbut the GRmouse strain (3) the

endogenously transmitted MMTV proviral

se-quences appear to have norolein early mam-marytumoroccurrence. One hypothesisis that these proviral sequences arise from the viral

infectionof the germcell ofthe animal (43). A

comparative study ofthe genomic locations of

the endogenous MMTV proviruses in different

inbred mouse strains may be informative in

de-termining their origins.

Mammary tumors arising in mice that were

exogenously infected by milk-borne virus have

been shownbyhybridizationanalysestocontain

not only the endogenous proviruses but also

additional proviral copies (22, 25). These are

assumedtohave arisen from theexogenousviral

infection ofmammary gland cells, followed by

reverse transcription and viral integration into

genomicDNA.These mammarytumor-specific

proviralsequences arethoughttoberelatedto

mammary tumorformation. The control of their

expressionmay bebased,in part,ontheprimary

nucleotidesequencewithin and around the

pro-viralintegration site.Therefore,wehavebegun

acharacterizationof theintegration sitesof the

exogenously acquired MMTV proviruses of

mammarytumorDNA.

We used the Southern DNA filter transfer

technique (37) tostudyboth the genomic

loca-tionsofendogenousMMTVprovirusesand the

genomicintegration sitesoftheexogenously

ac-quired provirusesinmammarytumorDNA.We

comparedtheendogenous provirusespresent in

the liverDNAof seven differentinbred mouse

strains. Six of the strains have been bred in the UnitedStates,and ofthese,two(CBAandC3H)

1013

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

1014 GRONER AND HYNES

canbedirectly traced from a BALB/candDBA mating (19, 38). The seventh strain, GR, has beenbred in Europe (26, 38). The resultssuggest

that the genomiclocations of the MMTV

pro-virusesin mouse strains of related geneticorigins

are similar. Conversely, the MMTV proviral

DNA in mousestrains of unrelated origins ap-pears to be located in different genomic sites. We determined theintegrationsites of the newly acquired MMTV proviruses in mammarytumor

DNAisolated fromindividualC3H orBALB/c mammary tumors. C3Hfemales produce milk-borne MMTV, and their offspring have a high incidence of mammary tumor occurrence. BALB/c mice transmit no milk-borne MMTV,

but when foster nursed on C3H mothers

(BALB/cfC3H), these animals also have a high

incidence of mammary tumor occurrence.

Ad-ditional MMTV-containing sequences were

identifiedin the mammary tumor DNA of both

strains. Theadditionalcopieswereintegrated at

differentgenomic sites in eachtumor.

MATERLALS AND METHODS Mouse strains and genetic origins. BALB/ cfC3H micebearingmammary tumorsandGR mice wereobtained from the Netherlands Cancer Institute, Amsterdam, The Netherlands. C3H/HeJ, BALB/c, CBA/H-T6,C57BL/6J, DBA/2J, and A/J micewere

obtained from colonies maintainedby the animalcare centerof the Swiss Institute forExperimentalCancer Research. All of the inbredstrains,with theexception

ofGR, have been bred in theUnited States. Approxi-mately60years agoaBagg albino femalewasmated with a DBA male. A number of different strains, includingC3H andCBA,canbe traceddirectlyfrom thismating. Inaddition,theBALB/cstrainoriginated

directly from the Bagg albinostock,and the A strain descended fromacrossofaBagg albino with another

mousestock. The C57BLstraincannotbetracedback directly from the above-mentionedcrossesbutit de-scends from stocks whichwereusedforbreedingother American strains (19, 38). Thus,five of these strains

areknowntobegenetically related,and thepossibility

that theC57BL strain is also relatedtoother Ameri-can-bred strains isnotunlikely.TheGR strain

origi-natedand has been bred inEurope(26, 38).

Purificationof MMTV viral RNA. GR cells from

a mousemammarytumorcelllineobtainedfrom K. Yamamoto weregrown in roller bottlesin Dulbecco-modifiedEaglemediumcontaining 10%fetal calf se-rumand 10-6M dexamethasone. The medium from confluent cultureswascollectedat24-hintervals and

clarifiedbycentrifugation,andthe viruswaspelleted byasecondcentrifugationfor45minat25,000 rpmin aBeckmanSW27rotor.The pelletedviruswas sus-pendedinTNE (0.1 MNaCl, 5mMEDTA,10mM

Tris-hydrochloride[pH7.5])andrepelletedina

Beck-manSW65rotorbycentrifugationfor30minat35,000

rpm. The viralpelletwasresuspendedinTNE,and proteinaseKandsodiumdodecylsulfate(SDS)were added to,respectively, 100,ug/ml and 0.5%. After

in-cubation for1hat370C,the viral RNA was extracted withphenol-chloroform and chloroform and precipi-tated by the addition of 2 volumes of ethanol. The RNA was dissolved in 10 mMTris-hydrochloride (pH 7.5)-5 mMEDTA and layered onto a sucrose gradient (10 to30%sucrose in 0.1MTris acetate[pH8.0],0.2% SDS) which was centrifuged for 2 h at 50,000 rpm at

20C

inaBeckman SW65 rotor. The 70S region of the gradient wascollected,theRNA wasprecipitated by theaddition of2volumes of ethanol, and the precipi-tate was collected bycentrifugation. The RNA was redissolved in 10 mM Tris-hydrochloride (pH 7.5)-5 mM EDTA and heat denatured for 2 min at

700C.

Thesolution was layered onto a second sucrose gra-dient madeasdescribed above and centrifuged for 4.5 hat50,000 rpm inaBeckman SW65 rotor. The RNA sedimenting in the 20 to 40S region was collected, precipitated by the addition of2volumes of ethanol, and isolated by centrifugation. The RNA was dis-solved in sterilewaterand storedat-20°C.

Isolation of DNA from mouse tissues. Mam-mary tumorsand organs were removed from the ani-mals, frozen inliquid nitrogen, and stored at-700C. One gram of tissuewasthawed in10mlof homogeni-zation buffer (0.3 M sucrose, 10 mM NaCl, 1.5 mM magnesium acetate, 10 mM Tris-hydrochloride [pH

8.0], 1 mM dithiothreitol, 0.1% Nonidet P-40) and

homogenized. The nuclei were sedimented by centrif-ugation for5min at3,000xg, thepelletwaswashed

3 times with 10 ml of homogenization buffer and suspended in 10 ml ofTNE, and proteinase K and SDSwereadded to,respectively,100,ug/ml and 0.5%. After incubation for 12 h at 370C, the DNA was extracted withphenol-chloroformandchloroform and

precipitatedby the addition of2volumes of ethanol. The DNA was collected by centrifugation and

sus-pendedin1mM EDTAto aconcentration of

approx-imately500,ug/ml.

Samples of DNA used in hybridization analyses

werebroughtto0.3NNaOH and incubatedat

1000C

for 10min todenature and decrease the size of the

DNA and to digest the nuclear RNA. The solution

wasneutralized, and the DNAwasprecipitatedby the addition of 2 volumes of ethanol and collected by

centrifugation. The pellet was dissolved in 1 mM EDTA, and the DNA concentrationwasdetermined by thediphenylamine analysis(6).

Preparation ofcDNA. MMTV-specific

comple-mentaryDNA(cDNA)wassynthesizedwith20 to40S

viralRNAasthetemplateinareaction mixture (0.1

ml) containing 0.05 M Tris-hydrochloride (pH 8.3);

0.05MKCI;8mMMgCl2;5mMdithiothreitol;1mM

dATP,1mMdGTP,1mMdTTP,0.1mM[3H]dCTP

(25.5Ci/mmol),or0.05mM[32P]dCTP(40Ci/mmol);

actinomycin D (100,ug/ml); RNA (15to 20,ug/ml); and avian myeloblastosisvirusreverse transcriptase

(100 U/ml).Thesynthesiswasprimed bydenatured fragments ofDNApreparedfrom Micrococcusluteus

DNA treated with DNase I asdescribed previously (40). TheDNAfragmentsat afinal concentration of

260 ,ug/ml were used to prime the synthesis. The

mixture wasincubated for2hat

250C,

andthe cDNA

wasisolatedasdescribedpreviously (15, 16).

For determining the extent towhich the MMTV RNAwasreversetranscribedintocDNA, increasing

J. VIROL.

http://jvi.asm.org/

(3)

amounts of cDNA were hybridized with a fixed amountof "25I-labeledMMTV RNAprepared as de-scribedpreviously (41).AtacDNA/RNAratio of2:1,

50% of thehybridizable RNAwas ina cDNA-RNA

hybrid, suggestingthatthe cDNA isafairly homoge-neouscopyof the templateRNA. Ata 10:1 cDNA/

RNAratio,amaximum of 80% of the RNA became

resistanttoRNase A(datanotshown).Thepurityof the cDNAwasassessedby analyzingforthepresence

ofsequencescomplementarytorRNA,themostlikely

contaminant of the viral RNApreparation. Withthe RNA ifiter transfer technique (1) total polysomal

RNA isolated fromaGR cell linewasboundtopaper

andhybridized with MMTV cDNA. There was no

evidence ofhybridizationin theregioncorresponding tomouserRNA(see Fig.3 ofreference12).

cDNAhybridizationto total cellular DNA. The number of MMTV proviral genomes in thevarious

preparations ofcellular DNA was determined by a

cDNAexcesshybridizationtechnique (4).TheDNA

(1.6to

10.ug)

waslyophilized togetherwithamounts of 3H-labeled MMTV cDNA(specific activity,19cpm/ pg) rangingfrom 30to 300 pgtoachieve theratios indicatedin thefigures.Thelyophilizednucleic acids

weredissolvedin 0.24 Msodiumphosphatebuffer(pH 6.8)containing1mM EDTAataconcentration of 0.5

mg/ml,denatured at 100°Cfor 10mm,andannealed underparaffinoil at69°Cfor 48 h.Thetotal content of each tubewasremovedinto 0.5 ml of SInuclease

assaybuffer(0.3MNaCl,0.03 Msodiumacetate[pH 4.5],3mMZnSO4, denaturedM.luteusDNA[10,g/

ml]) andthe proportion of cDNA which hybridized

wasdeterminedasdescribedpreviously (15, 16).

With thistechnique (4)it is alsopossibleto

deter-mine thekineticcomplexity ofthecDNA, providing

anadditional measure-forthe representativeness of theprobe.Formeasuringthekinetics ofhybridization, samples containing 1,000cpmofcDNAwereremoved at various time intervals into 0.5 ml of Si nuclease

assaybuffer, andtheproportionofhybridizedcDNA

wasdetermined.Figure1shows thekineticsof

hybrid-ization of MMTVcDNA with the total DNAisolated fromaC3Hmouseliver.Fromtheinitial slope of the curvetheDot1/2value of 6.9x 10-4ml-Iswas

calcu-lated. The substitution of this valueinto Bishop and

Freeman'sequation3(4) yieldedavalue forthekinetic

complexity of theMMTV cDNA. Afteracorrection wasmadefor the saltconcentration (5), this valuewas

foundto be 2.35 x 10', a number whichrepresents approximately80%of thereportedmolecular weight

of MMTVviralRNA(10).Theseresultsalsosuggest thatthe cDNAproberepresents mostof theMMTV genome.

Restriction endonuclease digestion of DNA.

Restriction endonucleases HindIII and BamHIwere

purchasedfrom NewEnglandBiolabs,Beverly,Mass., EcoRI was purchased from Boehringer Mannheim

Corp.,WestGermany,and PstIwasagift of G. Fey.

DNA (20,ug) wasdigestedwith atwofoldenzyme

excessin the buffersspecified bythemanufacturers.

Thedigestion mixtureswereincubatedfor6hat37°C

andthen foranadditional 30 minat370C with RNase

(50 ug/ml) toremoveRNA.Finally,the reactionwas

terminatedbytheadditionofSDSto0.2%. Toensure

complete digestion, in some cases the mixture was a

6-m

4 4-z a u

z

w

2-CL

z

-500 a

0 0

300 S

-100 a

Q

a.

4 8 12 16 20

HOURS

FIG. 1. Kinetics of hybridization of 3H-labeled MMTVcDNA with C3H liver DNA. C3H liver DNA

(15pg) washybridizedwith 1.27ngofMMTV

[3H]-cDNA in 30itlof hybridizationbufferat69°C.Atthe indicated times2-,ul samples were diluted into Si

nucleasedigestion buffer, and thepercentcDNA hy-bridized was determined as described in the text.

Approximately1%ofthe cDNAwasresistantto

diges-tionwithS1nuclease,and this valuewassubtracted

fromeachpoint.

extracted withchloroform,the DNAwasprecipitated

bythe addition of 2 volumes ofethanol,thepelletwas

resuspended in the appropriatebuffer,andthe DNA

wasredigestedwithatwofoldexcessofenzyme.After

the digestionreaction,eachsampleof DNAwas pre-cipitated bythe addition of 2 volumes ofethanol.The

precipitateswerecollected andsuspendedin 20pl of

10%Ficoll-0.1%SDS-1mM EDTA-0.02%

bromophe-nolblue.

Electrophoresis, transfer,andhybridizationof

enzyme-digestedDNA.Restrictionenzyme-digested

DNA samples were electrophoresed on 0.8 or 1.0%

agaroseslabgels (slotsurface,5by10mm)madein36 mM Tris-30 mM sodium phosphate (pH 7.2)-i mM EDTA at 18 V for 18 h at roomtemperature. The

digestedDNA and the molecularweightmarkerswere

visualizedbyethidiumbromidestainingand photog-raphyunder UVlight.The DNAfragmentsinthegel

weretransferredtoanitrocellulose filter sheetbythe

blotting technique ofSouthern (37) as modified by

Ketner andKelly (18).Beforehybridization,the

mem-brane filter (Millipore Corp., Bedford, Mass.) was

washedtwice in4xSSC(lx SSC is0.15MNaCl plus

0.015 M sodiumcitrate) plusDenhardtsolution(9)at 65°C,with the second washcontainingdenatured M. luteusDNA(20ug/ml) and 3.3%(vol/vol) phosphate

buffer (0.5 M sodiumphosphate, 1.5% sodium

pyro-phosphate [pH 7.3]). Thefilterwasthen hybridized

by saturatingit withasolution of32P-labeled MMTV cDNA (specific activity, 50 x 106 cpm//ug), SSC, EDTA, SDS, anddenatured M. luteusDNAsothat

the final concentrations were,respectively, 25ng/ml,

4x, 10 mM, 0.1%, and 20 uLg/ml. The hybridization

mixture and the filter were incubated in asealable

plasticbagbetweentwoglass platesfor 20 hat65°C.

The filterwas thenwashed at65°C with0.1%SDS and 3.3% (vol/vol) phosphate buffer containing 2x

SSC, followedbysecond and third washes in which the 2x SSCwasreplaced bylxSSC and thenby0.1x SSC foratotalperiodof3h. Thedryfilterwasthen

* 0

I

e.

http://jvi.asm.org/

[image:3.504.254.451.55.187.2]

(4)

exposed at-70°C to Kodak XR-5 film with an inten-sifying screen.

Preparation of and hybridization with nick-translatedmouse8-globin plasmid DNA. Plasmid pCR1containing themouse,B-globinstructural DNA sequences (34) was provided by B. Mach. The plasmid wasnick-translated with[32P]dCTP to a specific activ-ity of 3x107cpm/,ugasdescribed previously (46). The filters were prepared and hybridized as described abovewith theexception that the plasmid DNA was included at a final concentration of 100 ng/ml.

Isolation ofunintegrated MMTV viral DNA. HTC-M1 cells from a rat hepatoma cell line infected withMMTV(GR) (31) were obtained from K. Yama-moto. The cells were grown in Dulbecco-modified

Eagle medium, 10% fetal calf serum, and 10' dexa-methasone. The cells have been shown to contain unintegratedMMTV viral DNA (33) which was iso-lated as follows. Confluentcells wereremovedfrom

petridishes bytrypsinizationandwashed3times with cold phosphate-buffered saline. Approximately 109 cells weresuspendedin 50 ml of10mM

Tris-hydro-chloride(pH 8.0)-5 mM EDTA and disrupted witha

Potter homogenizer. The nuclei were collected by centrifugation, and thepelletwasreextracted in the samebuffer. Thesupernatantfractionscontaining the cytoplasmic, linear DNA molecules were pooled, and proteinase K andSDS were added to, respectively, 100 ,tg/ml and 0.5%. After extraction with phenol-chloroform and chloroform, the nucleic acids were precipitatedby the addition ofethanol,collected by centrifugation, and dissolved in2ml ofwater. Dena-tured M. luteus DNA (50 ,tg) was added, and the solutionwasbroughtto 0.1MNaCI-5mMEDTA-10 mM Tris-hydrochloride (pH 7.5). RNase, incubated for 10minat100°Cin 50 mM sodium citrate(pH5.0)

to removeDNase, wasaddedto 100,tg/ml, and the RNAwasdigested for2hat37°C.SDSwasaddedto

0.5%, and thedigestwasextracted with phenol-chlo-roform and chloroform. The aqueousphasewas

ap-pliedto aSephadexG-50column (0.5 by40cm)in10

mMTris-hydrochloride (pH 7.5),theexcluded volume waspooled,and thenucleic acidswereprecipitatedby the addition of2volumesofethanol. The DNAwas

collectedby centrifugationandsuspendedin 250

pi

of

1 mM EDTA. The content of MMTV-specific

se-quences was determined by hybridizing increasing

amountsof this solution withanexcessof 3H-labeled MMTV cDNA under the conditions described above. The hybridization was begun after incubating each reaction for10minat100°C.

RESULTS

MMTVproviral number inmouse DNA.

The DNA of all tested inbred strains ofmice

containsMMTVproviralsequences (22,25, 35,

44). Beforedoingthe genomicmappingstudies

of the MMTV proviruses of different mouse

strains, we determined, by hybridization, the

number of MMTV proviruses endogenous to

BALB/c,C3H, andGR mice.BALB/cmice do

nottransmitMMTV in themilkandhavealow

incidence ofmammarytumor occurrence. C3H

mice have ahigh incidence ofearlymammary

tumor occurrence and transmit viral particles in

the milk. In the GR mouse a single dominant

gene, the Mtv-2 locus, controls the high

inci-dence ofmammary tumor occurrence (3). This

strain isan exceptionas the endogenous

provi-rusesof other testedmouse strains seem to play

norole in earlymammary tumor occurrence. We

also determined, by hybridization, the number

of MMTVproviruses inC3HandGR mammary

tumors.

Otherinvestigators havedetermined the

num-ber of MMTV proviruses in organ DNA and

mammary tumor DNA isolated from various strains of mice by measuring the kinetics of

hybridization of radioactivecDNA to an excess

of cellular DNA (22, 25, 44). We employed a

cDNA excesshybridizationtechnique developed

by Bishop and Freeman (4) to determine the

amount of MMTV proviral genes in the DNA

isolated from livers and mammary tumors of

different mice. With this method both the

com-plexity of the cDNA probe and the comcom-plexity

of thehybridizable DNAcanbedetermined.

Thenumber of proviral genesinthe DNA of

eachmousestrainwascalculatedbydetermining the amount of cDNAhybridized at increasing

ratios of cDNAtoDNA. Thisvaluewas

repre-sentedaspicogramsof cDNAhybridper 100,tg

of DNA. The plateauvalue reached when the

cDNA-to-provirus ratio was about 10 to 20:1

reflects theamountofprovirusin the total

cel-lular DNA. This method has previously been

describedfordeterminingthe number ofglobin

genes in various cellular DNAs (4, 28). The

results of suchhybridizationsare showninFig.

2. Knowing thehybridizationplateauvalue and

the molecularweightofonegenomeequivalent

of MMTVproviralDNAandcorrectingfor the

fact that approximately 80% of the viral RNA

complexity is present in the cDNA (as

deter-mined above), we were able to calculate the

number of MMTVproviral copies. These

num-bers and the calculationsarepresentedinTable

1.C3H andBALB/cliver DNA andC3H

mam-mary gland DNA contained approximately 7

copiesof MMTVproviralDNAper

diploid

ge-nome, whereasC3HmammarytumorDNA

con-tained8.5 to 9copies. GR liverDNAand

mam-mary tumor DNA contained approximately 10

copiesof MMTVproviralDNAperdiploid

ge-nome. These data agree with previously

pub-lished work which shows that C3H mammary

tumors contain increased amounts of MMTV

proviralDNA(21, 22,25).

Restriction enzyme analysis of the

loca-tionof theendogenous MMTV

proviruses.

The digestionof the total genomic DNA with

restrictionenzymes,the separationof the

frag-ments by agarose gel electrophoresis, and the J. VIROL.

http://jvi.asm.org/

(5)

MMTV PROVIRAL DNA: GENOMIC NUMBER AND LOCATION

AB

0 ,

. . I I

0 00

@

4O.2-648-220 16 2

1017

4 8 12 16 20 4 8 12 16 20

ngcDNA/100pg DNA

FIG. 2. Saturationhybridization ofMMTV[3H]cDNAwithmouseDNA. DNA(1.6to10pg)waslyophilized

with amountsofMMTV[3HJcDNA ranging from30 to300pgtoachieve the indicated ratios. Thenuckeic acidsweresuspendedinhybridization bufferataDNAconcentrationof0.5mg/mLAfterincubationat69°C for48h, the entire mixturewasdilutedinto Si nuclease digestion buffer,andthepercent cDNAhybridized

wasdetermined.The countsperminutehybridizedwereconverted topicograms of hybridper100pgof input

DNA. The DNAwasisolatedfrom(A) C3Hliver(0), C3Hmammarygland (0),BALB/cliver(x); (B)C3H

mammarytumor;(C)GRliver; (D)GRmammarytumor.

TABLE 1. AmountofMMTV-specific DNA in mouse tissuesasdetermined byhybridization with MMTV

cDNA

pgof MT

cDNA MMTV

Mouse Tissue hybrid gene

nor

strain per 0

gene

nipo.d

1tg

of per diploid

DNA genome

C3H Liver 550 6.9a,6.3b

Mammary gland 550 6.9,6*0b Mammary tumor 680 8.6

BALB/c Liver 550 6.9

GR Liver 790 10.0

Mammary tumor 860 10.7

aDetermined as

follows:

the molecular weight of

oneproviral equivalentis6.0 x10'and one mouse cell has DNAwithamolecular weight of 3 x 1012;

there-fore, oneMMTV genome is 2 x 10-4% of the total genome. 100 ugof DNA containing one MMTV

ge-nomehas 200 pg of genebut only 100 pg of + strand. Aplateauvalue of550pg of cDNAhybrid per 100ug of DNA isequivalent to 687 pg of gene per 100

iLg

of DNA asthe cDNAprobe represents approximately

80%ofthe viralgenome. 687 pg of gene is equivalent

toapproximately6.9copies of MMTV per diploid cell.

bDetermined from the kineticsof MMTV cDNA

hybridization asdescribed in the text and in Fig. 2.

transfer of the DNA to anitrocellulosefilter (the Southern DNA filter transfer technique [37])

followed by hybridization with radioactive

cDNA allowed an analysis of the restriction

enzymesitesinproviralDNA and theflanking

host DNAsequences.We used thesetechniques

to analyzethe location of the MMTVprovirus in DNA isolatedfrom the livers andmammary tumorsof different inbred strains of mice. The

restriction enzymecuttingsites of the MMTV

viral DNAhave been determined.Unintegrated viral DNA hasbeen isolated from rathepatoma cells infected with MMTV and digested with

various restrictionenzymes(36;N. E.Hynes,B.

Groner, H. Diggelmann, R. Van Nie, and R.

Michalides, Cold SpringHarbor Symp. Quant.

Biol.,inpress).EcoRI andHindmcut the

pro-virus once close to the center of the genome,

whereas BamHI cuts it twice and yields an

internalfragmentofapproximately1.0kilobases

(kb) alongwith two endfragments.Upon diges-tion of the cellular DNA with EcoRI or with

HindIIl, each provirus yields two fragments

whosemolecularweights correspondto the dis-tancebetween the restrictionenzymesiteonthe

provirus and the nearest site in the cellular

DNA. This is also the case for digestion with

BamHI. An additional fragment of approxi-mately1.0kbcontaining only proviralDNAmay

alsoarise fromaBamHIdigestionof the

endog-enousproviruses.

For analyzing the genomic locations of the

endogenous MMTV proviruses, DNA isolated

from theliver,anorganwhich containsonlythe

endogenous proviruses, was digested with

EcoRI, HinduII, and BamHI. The DNA

frag-mentsresulting from the digestionofthe liver VOL. 33, 1980

800-

600-4c

z

a

400-co

0

200-a

0

=

800-z

4

Z

600-f

400-

http://jvi.asm.org/

[image:5.504.52.244.322.484.2]

(6)

DNAof various strainsof micewerefractionated

by electrophoresis and transferred to

nitrocel-lulose filters,andprovirus-containingfragments

wereidentified. Theresults of this analysisare

shown in Fig. 3.

For facilitating the comparison, the sizes of

theMMTV-containing EcoRIfragments of the

different strains are listed in Table 2. Such a

comparisonreveals thatamongtheseven

differ-ent strains, the BALB/c, CBA, and C57BL

strains had almost identical EcoRI fragments,

the C3H strain shared fragments with the

BALB/c mouse and the DBA mouse, andthe

DNA from the A mouse also had fragments commontothe BALB/cmouseDNAandtothe DNAofotherexaminedmice. Thesimilarityin

the EcoRIpatternsoftheCBA, BALB/c, and

Hind 3

C57BL DNA ledustodigestthese DNAs with

anotherenzyme. Lanes 2 and 4 of Fig.3reveal

thatthe MMTV-containing HindIII fragments

of CBA and BALB/c DNA were identical,

whereas the C57BL DNA yielded thesame

pat-ternwithtwoadditionalbands. Thepatternof

MMTV-containing EcoRI fragments arising

from theDNAof theGRmouseappearedtobe

different from that ofthe other strains tested.

Seven ofnine fragmentsidentifiedin Fig.3and

listed in Table 2 did not correspond to those

found in either the DBA DNAorBALB/c DNA.

The limits of resolution ofthe method donot

allow usatthis pointtoruleoutthepresenceof

additional EcoRIfragments,especiallythose of

thesize of about 7kb.This leavesthepossibility

thatadditionalEcoRIfragments ofsimilar

mo-Bam Hi Eco RI Eco Rl

1 2 Kb 3 4 5 Kb 6 7 8 9 10 11 Kb 12 13 14 15 16

...

-ivr-fteidctd resrit\'ion

bromide.ie.

... _~~~~~~~~~~~~~~~~~~~W5_ Q

[image:6.504.69.458.241.453.2]

7 ;

FIG. 3. AnalysisoftheendogenousMMTVprovirus in theDNAof differentstrainsofmice. DNA was isolatedfromtheliversoftheindicated strainsofmice anddigestedwithdifferentrestriction endonucleases.

Thesampleswereelectrophoresedinan0.8%agarosegel, transferredtoanitrocellulosefilter,andannealed

with32P-labeled MMTV cDNA asdescribed in the text. The markers indicate theposition in this and in

subsequentfigures ofthefragments ofEcoRI-cut lambdaphageDNAdetectedbystainingwith ethidium bromide.

TABLE 2. MMTV-containingDNA

fragmentsa

resultingfromtherestrictionenzymedigestion ofthe liver

DNAof differentmousestrains

Strain Enzyme Kboffragments

DBA/2J EcoRI 14.5 13.5 11.5 9.7 8.7 7.7 7.4 6.7 6.2 5.9 5.4 4.2

BALB/c EcoRI 14.5 9.6 7.5 7.15 6.7 6.1

C3H/HeJ EcoRI 14.5 7.5 6.7 6.2 5.9 5.4 4.2

CBA/H-T6 EcoRI 9.6 7.5 7.15 6.7 6.1

C57BL/6J EcoRI 10.3 9.6 7.7 7.5 7.15 6.7 6.1

A/J EcoRI 14.5 8.85 7.7 6.7 6.1 5.4 4.3

GR EcoRI 22.8 16.0 11.5 11.0 10.0 8.1 7.8 6.9 6.7

C3H/HeJ HindIII 22.0 16.5 12.0 8.2 7.0 6.0 4.3 4.15 3.7

C3H/HeJ BamHI 21.8 10.5 7.7 7.0 6.1 4.8 4.6 4.2

aOnlythefragmentslargerthan 3 kbarelisted.

J. VIROL.

http://jvi.asm.org/

[image:6.504.66.463.544.640.2]

(7)

MMTV PROVIRAL DNA: GENOMIC NUMBER AND LOCATION

lecular weightwere presentin the DNA ofGR

and DBA miceorofBALB/c mice.

Theresultscanbestbeinterpreted in lightof

the genetic origins of the different mice. As

previously noted,theC3HandCBA strainscan

betracedtothe cross of aBALB/cmousewith

aDBA mouse.EachMMTV-containingEcoRI

bandwascorrelated withaband presentinone

of the parents. The DNA from A mice, which

are related to the BALB/c strain, also had

MMTV bands in common with the BALB/c

DNAaswell aswiththeDNAofDBAandC3H

mice. The DNA from C57BL mice, which can

also betraced backtorelatedstocks,had, with

theexception oftwoadditionalbands, thesame

pattern of MMTV-containing fragmentsas the

BALB/cmouse.Theresultssuggestthatinmice

of relatedgeneticoriginsthepatternsof

MMTV-containing EcoRI fragments are similar,

sug-gesting thattheprovirusesarelocatedinsimilar

genomic environments. The pattern of

frag-ments arising from the GR mouse, which is

genetically unrelated to the others, appeared

different. Inaddition, the patterns of

MMTV-containing EcoRI fragments arising from two

other European bred strains (RIII and 020)

have also been identified and differ from each

other and from theGR strain (Hyneset al.,in

press).Thissuggeststhat the MMTVproviruses

in thesestrainsarelocatedindifferentgenomic

environments.

The interpretation of the preceding results

dependsupon two conditions. Onecondition is

thatthe integrated viral DNA is colinearwith

the unintegratedspecies. Thiscanbe testedat

least fornewlyintegratedmilkvirus-specific

pro-viral DNA. The resultssuggesting that this is

true arepresented below. The second condition

isthat whereastherestrictionenzymedigestion

patternof the viralDNAsynthesizedfrom

milk-borne MMTVisknown(36), the restriction

en-zyme digestion pattern for the MMTV

provi-ruses endogenous to different mouse strains is

notknown. Atthepresenttime thisproblem is

difficulttosolveasthere isnoreadilyavailable

sourceof viralRNA or DNAfromstrains, such

asBALB/c orC57BL/6J, withalow incidence

ofmammary tumor occurrence. Some evidence

forthe relatednessoftheendogenous proviruses

canbeobtainedfromthe resultsofthe BamHI

digestionpatterns. Figure 3, lanes 6 through 8,

shows that at least some of theproviruses

en-dogenousto GR, BALB/c, and C3H/HeJ mice

containedthe internalBamHI fragment of

ap-proximately1.0kb which also arose from a

diges-tion of the isolated viral DNA.However, another smaller fragment also arose from the BamHI

digestion ofGRaswell asC3H DNA. In

addi-tion, Cohen et al. (7) have published a pattern

ofMMTV-containing fragments resulting from

thedigestion of theBALB/c DNA with BamHI

in which the 1.0-kb fragment is absent. These

results suggest that these may be endogenous

proviruses with different BamHI sites. These

resultswill befurtherdiscussedbelow.

Analysis of the location of the

/8-globin

gene indifferent strains of mice.Inthe

pre-ceding section we analyzed the genomic

loca-tions ofthe MMTV provirusesin anumber of

differentmice. Weextended thiscomparisonto

another gene, that coding for

f3-globin.

Liver

DNA isolated from the different mice was

di-gested with EcoRI, an enzyme which doesnot

cutwithin the coding orintervening sequences

of the

fi-globin

gene (34, 42),and,followinggel

electrophoresis, the ,B-globin-specific fragments

wereidentified. The resultsareshowninFig.4.

C3H/HeJ, BALB/c, A/J, CBA/H-T6, and

DBA/2J mice exhibit the hemoglobin diffuse

gene product (38) and contain two linked

,B-globingenes(29).Lane 3shows thetwo

,B-globin-containing restrictionfragments of the BALB/c

DNAwhichmigrateatapproximately 14and7

kb, as previously identified (42). This pattern

appears identical to that obtained with DNA

isolated from the livers of CBA/H-T6, A/J,

Kb 1 2 3 4 5 6 7 8

WXc; w ;

Co > CcoI> MUl) -4

cr

~

0 co

z

[image:7.504.277.420.347.567.2]

m

FIG. 4. Analysisofthe

fl-globin

gene in the DNA

of different strains ofmice. DNA wasisolated from the liversof the indicated strains of mice, digested

withEcoRI,andanalyzedasdescribedin the legend toFig. 3, except that thefilters wereannealed with

32P-labelednick-translatedpCR1 plasmidDNA

con-taining ,B-globin-specific sequences. Thesizes of the

fragmentsinlanes 1 through 5 are 14.0 and7.0kb; in lanes 6through8thefragmentis10 kb.

1019 VOL. 33,1980

http://jvi.asm.org/

(8)

C3H/HeJ andDBA/2J mice (Fig. 4; lanes 1, 2,

4,and 5). TheC57BL/6J andGR mouse strains

contain only one

fl-globin

chain (38). Lanes 7

and 8 show that the

f,-globin

gene from both

strains was contained inan EcoRI DNA

frag-mentofapproximately10kb. Ourresults suggest

that in strainsof different genetic origins such

asGR and C57BL/6J themacroenvironmnentof

the

f,-globin

geneissimilar.

Analysis of MMTV proviral integration

sites in mammary tumor DNA. C3H mice

transmit MMTV in the milk and have a high

incidenceof earlymammary tumor occurrence.

BALB/c mice transmit no milk-borne MMTV

and have alow incidence oftumorformation.

However,whenBALB/c micearefosternursed

on C3H mothers (BALB/cfC3H), their

mam-mary tumor occurrencebecomes high.The

hy-bridization ofmammary tumorDNA has

dem-onstrated that there isanincreasein theamount

of MMTVproviruses. The extra copies are

as-sumed to derive from the exogenous-virus

ac-quired via the milk. It may be possible that

mammarytumorformation is relatedtothe site

ofintegration of the exogenous provirus in the

cellular DNA.Onecanpartiallycharacterize the

integration sites byanalyzing the DNA isolated

from differentmammary tumorswiththe DNA

filter transfer technique. The DNA from

individ-Bam Hi Eco Rl

ual C3H mammary tumors was isolated and

digestedwithdifferent restrictionenzymes.The

fragmentscontainingMMTV-specificsequences

weredetected, and the resultsareshowninFig.

5. It can be seen that the DNA of individual

C3Hmammary tumorscontained notonly the

restrictionfragmentspresentin the liver DNA,

but for each tumor it was possible to identify

additional MMTV fragments. Thesenewbands

areindicatedby the linesonthe left side of the

individual lanes. Itwas notpossible in allcases

toidentifynewMMTV-specific bands from the

digestion with one enzyme. This might have

beenduetothe coelectrophoresis of

tumor-spe-cificbands with those deriving from the

endog-enousproviralsequences.Inthe EcoRI digest of

tumor 3DNA (lane 6)nonewbandswereseen.

However, a new band was evident when the

same DNAwas digestedwithBamHI (lane 2).

If one copy of the provirus was additionally

integrated in the mammary tumor DNA, two

new MMTV-specific bands wereexpected. We

observed this for theEcoRI digestion oftumor

1DNA(lane8) and for theHindIIIdigestionof

tumor 4 DNA (lane 11). Other tumor-specific

bands in tumors 2, 3, and 5 might have been

maskedby theendogenousproviruses.

The DNAisolated from fourseparate tumors

removed fromoneBALB/cfC3Hmouse was

di-Hind 3

Kb 1 2 3 4 5 6 7 8 9 10 11

5

e-*P

,-;i

Kb 12 13

r.

4is -_

_E

w...*=_

. .

* ; t t A,

'. '* .

.' _

_w--'Xt' F

.*.

FIG. 5. AnalysisofMMTVproviralsequences in theDNAofindividual C3H mammarytumors.DNAwas isolated from five individual C3H mammary tumors (tl through t5), digested with different restriction endonucleases, andanalyzedasdescribed in thelegendtoFig.3.The markerstotheleft oflanes2, 3,5, 7, 8, 9, 11, 12,and13indicatethepositionof tumor-specific MMTV-containing fragmentsidentifiedinthe DNA isolatedfromthedifferenttumors.

J. VIROL.

http://jvi.asm.org/

(9)

gested with EcoRI, andthe ing MMTV-specific sequen

These results are shown inF

to the right, the bandsarisir

nous proviruses are indica

whereas the tumor-specific

cated by dotted lines. As w

C3H

tumors, each

BALB/cI

mor contained additional M ies.

It should be noted that

DNA was obtained from t]

kept in the Netherlands

whereas the BALB/cDNAu

13 of Fig. 3 was isolated fron

of the Swiss Institute for

E:

Research. TheMMTV DNA

arising from the SwissBALB

ing from the DNA of the A colony. Cohen et al. (7) ha

pattern

of MMTV-containin ments generated from BAL missing the 6.7-kb band and

terns

shown in Fig. 6.

Possil

these results could be that the original BALB/c stock w

that the Amsterdam colony(

sequences or that the BALI

zerland gained an MMTV

observation has also been i

and Varmus (8), who compa

Kb

21.8

-

7.5-5.9_

5.5-

4.8-

3.4-1 2 3 4

U."i

.,s__o

FIG. 6. Analysis of MMTVI the DNA of individual BALB/

mors. DNA isolated fromfour j

tumors removedfromone mou.

EcoRIand analyzed as descri

Fig. 3. In the schematic diagr bands arising from endogenous cated by solid lines, whereasti

specific fragmentsareindicatea

fragments contain- MMTV provirusesindifferent colonies ofC3H/

Ices were detected. St mice.

Pig.

6. In the scheme The results shown in Fig. 5 and 6 can be

ig from the endoge- summarized asfollows. First, although each

tu-Lted

by solid lines, mor contained at least one additional copy of

fragmnents are indi- MMTV proviral DNA, the extra bands in the

ras the case for the different tumor DNAs did not coelectrophorese.

FC3H

mammary tu- This suggests that exogenous proviralDNA can

[MTV proviral cop- integrate into differentsitesinmammarygland

cell DNA. Second, in most cases theadditional

the BALB/cfC3H bands in the tumor DNAs were clearly visible,

he BALB/c colony although they were not of thesameintensity as

Cancer Institute, those specific for the endogenous proviruses.

ised for lanes 10 and Therefore, it can be concluded thatahigh

per-amiceof the colony centage of the cells ineach tumorcontained the

Kperimental Cancer additional MMTV provirusesintegrated into the

v fragment of 6.7 kb same site in the genomic DNA. The results also

B/c

colony wasmiss- suggest that each tumor originatedfromasmall

msterdam

BALB/c number of cells or from one single

MMTV-in-ve also obtained a fected cell.

.g EcoRI DNA frag- A comparison of the genomic locations of the

.B/c

DNA which is newly acquiredproviral copies inmammary

tu-Iresembles the pat- mor DNA depends upon the assumption that

ble explanations for the integrated proviralDNAiscolinear with the

the proviral loci in unintegrated viral DNA. We tested this by

di-rere heterozygous or gesting the DNA isolated from a C3H mammary

of mice lost proviral tumor containing additionalMMTV proviral

se-B/c colony in Swit- quences with PstI, anenzyme which cleaves the

provirus. A similar proviral genome infiveplaces to yield six

frag-reported by Cohen ments of DNA which have beenorientated with

red the endogenous respect to the viral RNA genome (36). One end

fragment could not be detected because one

1 2 3 4 enzyme

recognition

sitewastoocloseto the end

of the proviral DNA. If the newly integrated

MMTV is colinear with the unintegrated viral

DNA, then the digestion of mammary tumor

DNA with PstI should yield bands which

cor-respond to the internal PstI cutting sites,

whereas the fragment which is joined to the

- -

=-

- cellular DNAshould have analtered molecular

weight.

The DNA froma C3H mammary tumor and

a C3H liver was digested with PstI, and the

fragments of DNA-containing MMTV

se-quences were identified. These were compared

to fragments arising from a PstI digestion of

unintegrated viral DNA isolated from

MMTV-infected cells. The results are shown in Fig. 7.

C3H mammary tumor DNA (lane 2) yielded

four fragments of 4.0, 1.7, 1.2, and 0.9 kb which

arose from internal PstI cutting siteson the viral

Droviral

sequences in DNA,whereas the

1.1-kb

fragment

arising from

.cfC3H mammary tu- the 3' end of the unintegrated MMTV viral

indiidual

mammary

DNA was notevident (compare lanes 1 and2).

se wasdigested with

'bed

in the legend to

This

suggests that the newly acquired mammary

am on the

rlght,

the tumor-specific provirus is attached to the

cellu-provirues

are

indi-

lar DNA somewhere within this end

fragment.

he additional tumor- Othermoredetailed restrictionenzymemapping

iby dottedlines. experiments ofnewly integrated mammary

tu-1021 VOL. 33,1980

http://jvi.asm.org/

[image:9.504.47.242.395.571.2]

(10)

1 Kb 2 3

[image:10.504.104.219.65.281.2]

§-4 -- F

FIG. 7. Analysis of MMTV(GR) viral DNA and C3HDNAdigestedwith PstI. Linear viral DNA (lane 1), C3H mammary tumor DNA (lane 2), and C3H liver DNA(lane 3) were digested with PstI and ana-lyzed asdescribed in the legend to Fig. 3, except that

a 1.0% agarosegel was used. The markers in the center indicatethe sizes of thefragmentsresulting

fromthePstI digestion ofMMTV(GR) viral DNA.

mor-specific proviruses (7) show that the

provi-rus is attached to the cellular DNA within a

region a few hundred base pairs from the ends

of the unintegrated viral DNA. These results

suggestthat newlyacquired proviruses are

co-linearly integrated.

Lane 2 reveals that C3H mammary tumor

DNAcontainedtwoadditionalfragments at 5.2

and 0.7 kb which strongly hybridized with

MMTV.These fragmentsarosefromthe

endog-enous proviruses as they were also present in

C3H liverDNA (lane 3). The intensityofthese

fragments suggests that they may have arisen

from intemalPstIsiteson theC3Hendogenous

proviruses. This leads to the conclusion that

C3Hendogenousproviruseshave differentPstI

cutting sites than doesviral DNA arising from

a milk-borne viral infection. In a recent

publi-cation (8),Cohen and Varmus havecome to a

similar conclusion fromastudyofthe

endoge-nous proviruses of the C3H strain and related

strains.

Digestion with PstI yielded fragments

com-prised of viral and cellular DNA andtherefore

shouldbe usefulin characterizing the genomic

location of proviral DNA. However, the bands

representing these fragments will be weaker than those representing the internal proviral

fragmentsastheyarisefrom only one provirus.

InFig. 7,lanes 2 and 3,itis possible to seefaint

bands which mightrepresent the endfragments,

but digestion with EcoRI, HindIII, or BamHI

yielded patterns whicharemore useful for

char-acterizing the genomiclocation of proviralDNA.

DISCUSSION

Inthispaper wepresentedthe results of

stud-ies of theMMTVproviralnumber and its

orga-nization within thecellularDNAisolatedfrom

organsand mammarytumorsof differentstrains

ofmice.

All strains of mice contain MMTV-specific

sequences intheirgenomicDNA(22, 25, 35, 44). By a cDNA excesshybridization technique (4),

wedetermined that C3H liverDNA,C3H

mam-marygland DNA,andBALB/cliver DNA

con-tainapproximately7copies each of MMTVper

diploidgenome,whereasGR liverDNAcontains

approximately 10. These numbers are in close

agreement with published values (22, 25). As

previously observed (22, 25),wefound that C3H

mammary tumor DNAcontains additional

cop-ies of the MMTVprovirus.

We used the DNA filter transfer technique

(37) to compare the locations of the MMTV proviruses endogenous to different mouse strains.The results reveal thatin mousestrains

ofunrelatedorigins, suchasBALB/c,

DBA/H-T6, andGR,the patternsofMMTV-containing

restriction enzyme fragments arising from the

digestion of liverDNAs are different.In mouse

strains ofrelated genetic background the pat-terns ofMMTV-specific fragmentsare similar.

One interpretation of these results is that the

endogenous MMTV proviruses of unrelated

mousestrainsareindifferentgenomiclocations,

whereas inrelatedmice theirgenomic locations

aresimilar. Onehypothesiswhichcouldexplain

the results is that proviral DNA arose from a

number of independent germ cell infection

eventsin which viralDNAwasabletointegrate

intodifferentgenomic locations.However,once

the viral DNA was integrated it remained a

stablegeneticelement.Hence,we areabletosee

similarities in thegenomiclocations of the pro-viruses ofgenetically related mice. Cohen and

Varmus (8) have recently published a similar

analysis and have come to similar conclusions.

They also saw that different inbred mouse

strainsresulting from the Baggalbino x DBA

matingcarried MMTVproviruses insimilar

ge-nomic locations. In addition, they determined

that thegenomiclocations of MMTVproviruses

inferal micearedifferent, resultswhichcanbe

compared to our findings on inbred mouse

strains which are genetically unrelated. In

an-J. VIROL.

http://jvi.asm.org/

(11)

otherpublication (24), the authors determined

that the MMTVproviruses endogenous tothe

A/HeJmouse arelocatedon atleast three

sep-aratechromosomalpairs. This result alsofavors

theidea thatseparate viral germ line infection

events ledtoproviral DNA. Analternative

in-terpretation of the differences seen inthe

pat-terns of MMTV-containing fragments in, for

example, the GR and C57BL DNA is that the

sequencessurrounding theprovirusesinthe

dif-ferentmice have diverged. To analyze this

pos-sibilityweused theSouthem DNA filter transfer

technique (37) to analyze the/3-globin gene in

the different mousestrains. C57BL/6J and GR

micehave theHb-S locus and code for onlyone

f8-globinprotein. Both strains contain the

,8-glo-bingenewithina10-kb EcoRI fragmentofDNA.

Recently ithasbeenreported that theC57BL/

10 mouse has its fl-globin gene within a 10-kb

EcoRI fragment (45). Upon further analysisby

RPC5chromatography, thisfragmentwasfound

to be composed oftwo different 10-kb pieces

whichcould bedifferentiated by digestion with

restriction enzymes. Whether the 10-kb band

arising from the digestion of GR DNA with

EcoRI also contains twoseparate fragments is

not known. The 83-globin gene in these two

mousestrains ofdifferentgeneticoriginsappears

tobein asimilar macroenvironment. Thesame

analysis has been carriedoutfor the a-amylase

gene. Arecombinantplasmid containing

a-am-ylase mRNA-specific sequences recognizes at

least four genomicDNAEcoRIfragments.The

DNA from the seven different mouse strains

whichweanalyzedgaverisetothesame pattern

of a-amylase-containing EcoRI fragments (B.

Groner, N. E. Hynes, and U. Schibler,

unpub-lishedobservations). These resultssuggestthat

these structural genes in the different mouse

strains aresimilarlyorganized,whereas MMTV

proviral DNA is located in different genomic

sites. This couldreflectseparate germ line

inte-gration events into a relatively similar

back-ground of structural genes in the different

strains. The validityofthis idea awaits further

comparative studies with probes specific for

otherstructural genes.

One difficulty in comparing the endogenous

proviruses ofdifferent strains is that the

restric-tion enzyme sites in these proviruses have not been mapped and may vary from the pattern

determinedforisolated viral DNA. For most of

themousestrains itappears that theendogenous

proviruses never yield viral particles. Thus,

pu-rified genomic DNA may be the best substrate

for suchananalysis. Previouspublications sug-gest thatthereis a highdegree of cross hybrid-izationbetween the endogenous MMTV

provi-ruses of different strains (20, 30). However,

RNase T1 analyses (11) andhybridization

anal-yses (20) have shown that differences do exist.

Aspointedoutin theresults,wehave evidence

that atleastsomeof the restrictionenzymesites

ofendogenousproviruses correspondtosites in

isolated viral DNA. The viral internal BamHI

fragment of 1.0 kb also appears in a BamHI

digestion ofliverDNA fromC3H, BALB/c, and

GR mice. However, the digestion ofC3H liver

DNA withPstIreveals thatonlytwoof the four

internalPstI fragments obtained from a digest

of viral DNAareidentical. This resultsuggests

thatsomeof the restrictionenzyme sites of the

endogenous proviruses vary from those

deter-mined for isolated viral DNA.

An additional observation which we have

made in the courseofthesecomparativestudies

is that DNAisolated from different coloniesof

BALB/c mice yields different numbers of

MMTV-containing EcoRI fragments. There is

oneadditionalfragmentpresentin the DNA of

the Swisscolony. Weattemptedtoanalyzethis

difference further by digesting the DNA with

PstI. This enzyme has been used in this work

(Fig.7)and in others (7,8)todistinguish

endog-enous MMTV proviruses from those

corre-sponding to the exogenous milk-bome virus.

Upon digestion with PstI, DNA isolated from

mice of both colonies gave rise to the same

pattern (datanotshown),whichhaspreviously

beenpublished (8).Ifthe SwissBALB/ccolony

hasacquired a newMMTV provirus,itcannot

be distinguished from the others by PstI and

therefore does not correspond to the BALB/

cfC3Htumor-specific viral variant.

With theSouthemblottingtechniquewe

an-alyzed theDNAofmammary tumorstaken from

C3H miceorfromBALB/c mice foster nursed

on C3H mothers. The animals received exoge-nous, milk-borne MMTV, and we determined

that each tumor contained at least one

addi-tionalcopy of MMTVproviralDNA.It is

inter-esting to note thattumors fromBALB/cfC3H

animals contain higher amounts of newly

ac-quiredproviruses than those isolated from C3H

animals. We analyzed a number ofC3H

mam-mary tumorsand have detectedamaximum of

one new proviral copy, whereas one of the

BALB/cfC3HtumorDNAsshown inFig.6

con-tained at least four newproviral copies. Other

investigators (7) who looked at BALB/cfC3H

tumorsalso found highamounts ofnewly

inte-grated MMTV proviruses. Perhaps C3H mice

are more resistant to the C3H virus than are

BALB/c mice. Different types ofexperiments

haveled others to suggest this (13).

The comparison of the pattern of MMTV-33,

http://jvi.asm.org/

(12)

containing restriction enzyme fragments

ob-tained with different tumor DNAs allows the

following conclusionsto be made. First, the

pat-temof the extra MMTV-containing DNA bands

is different in each tumor, suggesting that the

exogenous viral DNAcanintegrate at different

sites in the hostgenome. Second, the fact that

extrabands ofMMTV-containing DNA are ev-ident suggests that a high percentage of the

tumorcells containedthe exogenous viral DNA

integratedin thesame site and that the tumor

mayhave arisen fromone cell or from a small

number ofinfected mammary gland cells.

Intwo recentpublications(7, 32), other

inves-tigators have reported that MMTV proviral

DNA can integrate into a number ofgenomic

sites. In one case (7), the integration sites of

newly acquired, tumor-specific proviruses were

compared.Inthe otherreport(32), the genomic

locations of the proviral DNA in an

MMTV-infectedratcell line were determined. The

ex-ogenous infectionof cells with otherretroviruses,

including murine leukemia virus (39), spleen

ne-crosis virus (17), and aviansarcoma virus (14),

has yielded results which suggest that the

pro-viral DNAspecific for these virusescanintegrate

intoanumber ofgenomic sites.

The differentpatterns of MMTV-containing

sequences invariousmousestrainsand the

dif-ferentproviral integrationsitesfoundin

individ-ual mammary tumor DNAssuggestthat there

aremultiplesites onthe cellular genome

avail-ablefor proviral DNA integration. Conclusive

evidence as to whether the different MMTV

integration sites sharecommonfeaturescanonly

be drawn from a sequence analysis of the

ge-nomic DNAflanking the provirus. Suchan

anal-ysis may help to clarify the mode of proviral

DNAintegration andmayrevealsequences

re-latedtothe controlofMMTVproviral

expres-sion.

ACKNOWLEDGMENTS

The support and encouragement of H. Diggelmann, in whoselaboratory this workwasperformed, isgratefully ac-knowledged.Wethank J.Beard for the avianmyeloblastosis virus reversetranscriptase,B.Mach forplasmid pCR1DNA containingthe8-globinsequences,L. Bootand J.Hilgersfor supplyinguswithsomeof the mammarytumor-bearing mice, G.Fey for PstI, and K. Yamamoto for the GR cell line. We also thankR.Michalides forcritically readingthemanuscript. Finally wethank S. CherpillodandP. Dubied forhelpin preparing the manuscript.

Thisworkwassupported by grants 3.706-0.76SR and 3.273-0.78SR fromtheSwiss NationalScience Foundation. N.E.H. wassupported by a grant from the Swiss Cancer League.

LITERATURE CITED

1. Alwine,J.C., D. Kemp, and G. R. Stark.1977.Method for detectionofspecificRNA inagarosegelsby transfer todiazobenzyloxymethyl-paper and hybridizationwith DNA probes.Proc. Natl.Acad. Sci. U.S.A. 74:5350-5354.

2. Bentvelzen,P. 1974. Host-virus interactions in murine

mammarycarcinogenesis.Biochim. Biophys.Acta 355: 236-259.

3. Bentvelzen, P., and J. H. Daams. 1969. Hereditary infections with mammary tumor viruses in mice. J. Natl. Cancer Inst. 43:1025-1035.

4.Bishop, J. O., and K. B. Freeman. 1974. DNA sequences neighboring the duck hemoglobin genes. Cold Spring Harbor Symp. Quant. Biol. 38:707-716.

5. Britten, R. J., D. E. Graham, and B. R. Neufeld. 1974. Analysis of repeating DNA sequences by reassociation. Methods Enzymol. 29:363-418.

6. Burton, K. 1968. Determination of DNA concentration withdiphenylamine. MethodsEnzymol. 12B:163-166. 7. Cohen,J.C.,P.R. Shank, V.L. Morris, R. Cardiff, and H. E. Varmus. 1979. Integration of the DNA of mousemammary tumorvirus in virus-infectednormal andneoplastic tissue of the mouse. Cell 16:333-345. 8. Cohen,J. C., and H. E. Varmus. 1979. Endogenous

mammarytumor virusDNA varies among wild mice and segregates during inbreeding. Nature (London) 278:418 423.

9. Denhardt, D. T. 1966. A membrane-filter technique for the detection of complementary DNA. Biochem. Bio-phys. Res. Commun. 23:641-646.

10. Dion,A.S.,U.L.Heine,A.A. Pomenti, J. Korb, and G. H. Weber. 1977. Electrophoretic analysis of the molecular weight of murine mammary tumor virus RNA. J. Virol. 22:822-825.

11. Friedrich, R.,V. L.Morris,H. M.Goodman, J. M. Bishop, and H. E. Varmus. 1976. Differences between genomesof twostrainsof mouse mammary tumor virus asshownby partial RNA sequence analysis. Virology 72:330-340.

12. Groner, B.,N. E.Hynes,and H.Diggelmann. 1979. Identification of mouse mammary tumor virus-specific mRNA. J.Virol. 30:417-420.

13. Hilgers, J., and P.Bentvelzen. 1978. Interaction be-tweenviral and genetic factors in murine mammary cancer.Adv.Cancer Res. 26:143-195.

14. Hughes,S.H.,P. R.Shank,D.H.Spector,H. J.Kung, J. M.Bishop,H. E. Varmus, P. K. Vogt, and M. L. Breitman.1978.Proviruses of aviansarcomavirus are terminally redundant, co-extensive with unintegrated linearDNAandintegrated at many sites. Cell 15:1397-1410.

15. Hynes, N.E., B. Groner,A. E. Sippel, S. Jeep,T. Wurtz, M. C. Nguyen-Huu, K.Giesecke, and G. Schiitz.1979.Control of cellularcontentofchicken egg whiteprotein-specificRNAduringestrogen administra-tionand withdrawal.Biochemistry18:616-624. 16. Hynes,N.E.,B.Groner,A. E.Sippel,M. C.

Nguyen-Huu,andG.Schutz.1977.mRNAcomplexityand egg whiteproteinmRNAcontentinmatureand hormone-withdrawn oviduct.Cell11:923-932.

17.Keshet, E.,and H. M. Temin. 1978. Sites ofintegration ofreticuloendotheliosis virusDNA in chicken DNA. Proc.Natl. Acad. Sci. U.S.A. 75:3372-3376.

18.Ketner, G., andT. J. Kelly. 1976. Integrated simian virus40sequencesintransformed cell DNA: analysis

usingrestriction endonucleases. Proc. Natl. Acad. Sci. U.S.A. 73:1102-1106.

19. Klein,J. 1975.Biologyof themousehistocompatibility-2

complex: principles of immunogenetics applied to a

single system, p.26. Springer-VerlagNewYork, Inc., New York.

20. Michalides, R., and J. Schlom. 1975.Relationshipin nucleic acid sequencesbetweenmousemammarytumor

virusvariants.Proc.Natl. Acad.Sci. U.S.A. 72:4635-4639.

21. Michalides, R.,L.VanDeemter,R.Nusse,G.Ropcke,

and L. Boot. 1978.Involvement ofmousemammary tumor virus in spontaneous and hormone-induced mammary tumors in low-mammary-tumor mouse

strains.J. Virol. 27:551-559.

J. VIROL.

http://jvi.asm.org/

(13)

22.Michalides, R.,G.Vlahakis, and J.Schlom.1976. A biochemical approach to the study of the transmission of mousemammary tumor viruses in mouse strainsRu andC3H. Int.J. Cancer 18:105-115.

23. Moore, D. H. 1975.Mammary tumor virus, p. 131-167. In F.F. Becker(ed.), Cancer:acomprehensive treatise, vol. 2. Plenum Publishing Corp., New York.

24. Morris, V. L, C.Kozak, J. C. Cohen, P. R. Shank, P. Jolicoeur,F.Ruddle,and H. E. Varmus. 1979. En-dogenous mouse mammary tumor virus DNA is distrib-uted amongmultiplemousechromosomes.Virology 92: 46-55.

25.Morris, V. L., E. Medeiros, G. M. Ringold, J. M. Bishop, and H. E. Varmus. 1977. Comparison of mousemammarytumorvirus-specific DNA in inbred, wild and Asian mice, and in tumors and normal organs from inbred mice. J. Mol. Biol. 114:73-91.

26.Miihlbock,0. 1965. Note on a new inbred mouse strain GR/A. Eur. J. Cancer 1:123-124.

27. Nandi, S.,andC. M. McGrath. 1973.Mammary neopla-sia in mice.Adv. Cancer Res. 17:353-414.

28. Old, J., J. B.Clegg,D. J.Weatherall, S. Ottelenghi, P. Comi,B. Giglioni,J.Mitchell, P. Tolstoshev, and R. Williamson. 1976. A directestimateof the number ofhuman,B-globin genes. Cell 8:13-18. 29. Popp,R.A.,and E. G. Bailiff.1973.Sequenceof amino

acids in themajor and minor , chains of the diffuse hemoglobin from BALB/c mice. Biochim. Biophys. Acta 303:61-67.

30.Ringold, G. M.,P. B.Blair,J. M.Bishop,and H. E. Varmus. 1976. Nucleotidesequence homologies among mousemammary tumor viruses.Virology 70:550-553. 31.Ringold,G.M.,R. D.Cardiff,H. E.Varmus,and K. R. Yamamoto. 1977. Infection ofcultured rat hepa-tomacellsby mouse mammary tumor virus.Cell 10:11-18.

32. Ringold, G.M., P. R.Shank,H. E.Varmus, J. Ring, and K. R. Yamamoto. 1979.Integration and transcrip-tionofmousemammarytumor virus DNA in rat hep-atomacells. Proc. Natl.Acad. Sci.U.S.A. 76:665-669. 33.Ringold, G. M.,K. R.Yamamoto, P. R. Shank, and H. E. Varmus.1977.Mouse mammary tumor virus DNA in infected rat cells: characterization of unintegrated forms.Cell 10:19-26.

34. Rougeon,F.,and B. Mach. 1977. Cloning and amplifi-cationof a and,8mouse globingene sequences

synthe-sized in vitro. Gene 1:229-239.

35.Scolnick, E. M., W. Parks, T.Kawakami,D.Kohne, H.Okabe, R. Gilden, and M. Hatanaka. 1974. Pri-mateandmurinetype-Cviral nucleic acid association kinetics:analysis ofmodel systems and natural tissues. J. Virol. 13:363-369.

36.Shank, P. R., J. C. Cohen, H. E. Varmus, K. R. Yamamoto, and G. M. Ringold. 1978. Mapping of linear and circular forms of mousemammary tumor virus DNA with restrictionendonucleases: evidence for alarge specific deletion occurring athigh frequency during circularization. Proc. Natl.Acad. Sci. U.S.A. 75: 2112-2116.

37. Southern, E. M. 1975. Detection ofspecific sequences among DNAfragments separatedby gel electrophore-sis. J. Mol. Biol. 98:503-517.

38.Staats, J. 1976. Standardizednomenclature for inbred strains of mice:sixthlisting.Cancer Res. 36:4333-4377. 39.Steffen,D., and R. A.Weinberg. 1978.Theintegrated genome of murine leukemia virus. Cell 15:1003-1010. 40.Taylor,J.M., R.Illmensee, and J.Summers. 1976. Efficient transcription of RNA into DNA by avian sarcomaviruspolymerase.Biochim.Biophys.Acta 442: 324-330.

41. Tereba, A.,and B. J.McCarthy.1973.Hybridizationof 5I-labeled RNA.Biochemistry12:4675-4679. 42. Tiemeier,D.C.,S. M.Tilghman,F. I.Polsky,J.G.

Seidman,A. Leder,M.H. Edgell, and P.Leder. 1978.Acomparisonof two cloned mouse,B-globingenes and theirsurroundingandinterveningsequences. Cell 14:237-245.

43. Todaro, G.J., R. E.Benveniste,R.Callahan,M. M. Leiber, andC.J.Sherr.1975. Endogenous primate and felinetype-C viruses. ColdSpringHarborSymp. Quant. Biol.39:1159-1168.

44. Varmus, H. E., J. M.Bishop,R.C.Nowinski,and N. H. Sarkar. 1972.Mammary tumor virus specific nu-cleotide sequences in mouseDNA. Nature (London) NewBiol. 238:189-191.

45. Weaver, S.,N.L.Haigwood,C. A.Hutchison,and M. H.Edgell.1979.DNAfragmentsof theMus musculus ,B-globin haplotypesHbb' and Hbbd. Proc.Natl.Acad. Sci.U.S.A.76:1385-1389.

46. Weinstock,R., R.Sweet,M.Weiss,H.Cedar,and R. Axel.1978.IntragenicDNAspacersinterruptthe oval-bumin gene. Proc. Natl. Acad. Sci.U.S.A. 75:1299-1303.

http://jvi.asm.org/