Purification and characterization of the DNA polymerase of human

MedicalSciences

Purification

and

characterization

of the

DNA

polymerase

of

human

breast

cancer

particles

(reversetranscriptase/humanmalignancy/oncornavirus/synthetictemplates)

T. OHNO, R. W. SWEET, R. Hu, D.

DEJAK,

ANDS. SPIEGELMAN

InstituteofCancerResearch,College of Physicians andSurgeonsofColumbiaUniversity, 701 West 168 Street, NewYork,N.Y.10032

Contribted byS.Spiegelman, November29,1976

ABSTRACr Previous studies haveidentified human breast tumorparticles possessing many of the features characteristic of RNAtumorviruses. Inadditiontothe expected size (600 S) and density (1.16,g/ml) these include possession ofanouter

membraneandaninneronesurrounding a "core" containing

a DNApolymerase anda-large-molecular-weight (70 S) RNA possessing detectable homology to the RNAs of the mouse mammary tumorvirus(MMTV) and of the Mason-Pfizer

mon-keyvirus (MPMV).

Wereport here the purification and characterization of the

DNApolymerasefromthe human breastcancerparticles.'Its key propertiesareverysimilartothoseoftheRNA-dependent DNAnucleotidyltransferase(reverse transcriptases) foundin

MMTVand MPMV.Thus,likethese viralenzymes, the purified humanbreast cancerDNApolymeraseexhibitsthefollowing three features thattogether

distinguish

theknown viralreverse

transcriptases from normal' cellular DNApolymerases: (i) a strong preference foroligo(dT)poly(rA)overoligo(dT)poly(dA)

as -atemplate for thesynthesisof

poly(dT);

(ii)theacceptance

of thehighlyspecificoligo(dG)

py(rCm)

as atemplate

for

the formation ofpoly(dG) (iii)the

ability

touse aviralRNA(AMV)

as atemplatetofashionafaithfulDNAcomplementarycopy; and (iv) itspreferenceforMg++overMn++.

In summary, the data described hereontheenzymeof the human breast-cancerparticles add further evidence of simi-laritiestotheviral agents associated with thecorresponding malignancies inthemouse andmonkeymodels. Todate,an enzyme with these properties hasnotbeen detectedinnormal

breasttissuesor inbenigntumorsof the breast.

The first molecular indications of virus-like components in

human breastcanceremerged (1)fromhybridizationsbetween

the RNAsofbreasttumorsand

[3H]DNA

complementary

to the RNA of themousemammarytumorvirus

(MMTV),

find-ings thatwere

independently

confirmed

(2).

Another agent yieldingaprobe

(3)

thatcandetect

homologous

RNAinhuman

breastmalignanciesisthe Mason-Pfizer

monkey

virus(MPMV),

an isolate from a spontaneous breast carcinomaof a rhesus

monkey (4). Neitherof these

homologies

weredetectedinthe RNAsof normaltissues orof

benign

tumors

(fibroadenomas)

of the breast(1, 3).

Tofurther delineatethe nature ofthe viral-related RNA

sequencesfoundinbreast

malignancies,

recourse washadto

the"simultaneousdetection test"ofSchlomandSpiegelman (5). Thisexperimental procedurewas

designed

todetectthe

presenceofalarge (70Sor35

S)

RNAmolecule

complexed

to aDNA

polymerase

(areversetranscriptase, RNA-dependent

DNA

nucleotidyltransferase)

thatcanusetheRNAas a

tem-platetomakea DNA

complementary

copy. It wasfound

(6)

Abbreviations: MMTV, murine mammary tumor virus; MPMV,

Mason-Pfizermonkeyvirus;oligo(dG)-poly(rCm), oligodeoxyguany-late-poly(2'-O.methylcytidylate);TNEbuffer,buffer containing0.01 MTris-HCG atpH8.0,Q.15 MNaCl,3mMEDTA;DTT,dithiothreitol; solutionA,solutioncontaining2,mMDTT,1 mMEDTA,0.02% Triton X-100(wt/vol), 10% glycerol (vol/vol);AMV,avianmyeloblastosis

virus.

that 79%of-the malignant specimens examined by thistest

showedpositiveevidence of such RNA-enzymecomplexes.In contrast, noneofthe control samples from normal tissueand benigntumorswerepositive. It wasfurther establishedinthe

same study that the RNA-reverse transcriptase complexes identifiedinthe malignanttissues wereencapsulated in parti-clespossessing adensity between1.16and1.19g/ml,avalue characteristic ofintactanimal RNA tumor viruses. Finally,it wasshown (7) that these human tumor particles, likethose

foundinmousemammary tumors, can beconverted by'Ster-ox-SL treatment into "coreparticulates" with a density of1.26

g/mlorgreaterand containing thediagnostic reverse tran-scriptase and 70S RNA. Recently, the presence of reverse

transcriptaseinhuman breast cancer tissue wasestablished(8)

by the use of the highly specific synthetic template,

oligo-deoxyguanylate-poly(2'-O-methylcytidylate)

[oligo(dG)-po-ly(rCm)].

It isevident that the particles foundinhuman breastcancers exhibitmany of thefeaturescharacteristic of the twoanimal

oncornaviruses(MMTV and MPMV) associated with the cor-responding disease in mice and monkeys. Thepotential

im-plications of their presence stimulated further attempts at characterizingthecomponents ofthebreast cancer particles.

Thereversetranscriptase they contain is particularly amenable

toisolation since it canbeeasily detected during fractionation byarelatively sensitive assay of its activity. Indeed we have already demonstrated (9) thefeasibility ofthisapproach by purifyingtohomogeneity the reverse transcriptase from

par-ticles found in thespleensofleukemic patients.

It isthe purposeofthe present paper to describe the

prop-ertiesof the reversetranscriptase purified fromthe human breast cancerparticles. We find thatthe'human enzyme is very similarinitsdistinctive properties tothe known animal reverse transcriptases.

MATERIALS AND METHODS

Subcellular Fractionation ofBreast Tumor Tissue.

De-pendingontheamountsofmaterialavailable, between9and

30gof tumor werethawed, minced, suspendedin 4volumes of cold 5% sucrose(wt/vol)-TNE(0.01MTris-HClatpH 8.0,

0.15 M NaCl,3 mMEDTA) and blendedin aSilverson ho-mogenizer. Thehomogenatewascentrifugedat 4000Xgand then10,000Xgtoremovenuclei andmitochondria, respec-tively. Trypsin(WorthingtonBiochemicals)wasadded to the

postmitochondrialsupernatantto afinalconcentrationof0.5

mg/ml.After incubationat200for 10mmi,proteolyticactivity

wasinhibitedby the additionof limabean trypsininhibitor

(1-fold excess,WorthingtonBiochemicals)andTrasylol (100 kallikreininactivatorunits/ml,Mobay ChemicalCorp). The

sample was layered over-discontinuous sucrose gradients composed of6mlof 50o sucrose(wt/vol-TNEand8ml of25%

sucrose(wt/vol)-TNE.After

centrifugation

at25,000rpmfor 764

90 min at40 inaSpinco SW-27rotor,materialatthe

25/50%

interface wascollected, diluted withTNE, and layeredover

linear 20-50% sucrose-TNE gradients. The samples were

centrifuged asabove for 16hrand the different densityregions

collected. Thedensityregion(1.16-1.19g/cm3)inwhichRNA

tumorviruseslocalizewaspooled, diluted, andcentrifugedas

above for 90 min. The resulting pelletswere resuspended in

approximately 0.6mlof0.1 MTris-HClatpH 8.0.

Sixlots oftumorswereprocessedforenzymeinthemanner

described. Fourof these(A,B, C,andD)yielded enough

ma-terial to characterize the enzyme. One preparation (A), a

metastaticlivertumor,camefromasinglepatient;all theothers

being pooled material fromanumber of different

individu-als.

PolyacrylamideAgaroseGel Filtration. Theresuspended pelletwassolubilizedand disruptedat0° for 15 min by the

additionofKCI (to0.4M),DTT(dithiothreitol,to0.01 M)and TritonX-100(to0.6%, vol/vol). The sample, approximately0.9

ml,wasapplied toa0.9 X50cmcolumnofpolyacrylamide

agarosegel (Ultrogel AcA44, LKB) equilibrated with 0.3 M

potassiumphosphateatpH 8.0insolutionA(2mM DTT, 1mM

EDTA, 0.02% Triton X-100, and 10% glycerol). The column

waselutedwiththesamebufferataflowrateofabout2ml/hr.

Fractions(0.5ml)wereassayed forDNApolymerase and

ter-minaltransferaseactivitiesasdescribed below.

Phosphocellulose Chromatography. The peak fractions fromtheUltrogel columnwerepooled(3ml)and Trasylolwas

addedtoaconcentration of 100kallikreininactivatorunits/ml.

Thesamplewasdialyzedagainst0.01 Mpotassiumphosphate

atpH 7.2,insolutionAuntil thephosphateconcentration was

less than 0.02 M, and then was loaded onto a 0.9 X 10 cm

phosphocellulose column(Whatman P-ll) equilibratedwith

thesamebuffer.Thecolumnwaswashed with30mlofthe0.01

Mphosphatebufferand theenzymeactivitywaseluted with

a 120 ml linear gradientof0.01-0.5 M potassium phosphate

buffer at pH 7.2 in solution A, at a flow rate of 14 ml/hr. Fractions(1.2ml)wereassayed for bothDNApolymeraseand

terminal transferase activities. The fractions containing the mainpeakofpolymeraseactivitywerepooledand Trasylolwas

added to 100 kallikrein inactivator units/ml. This enzyme

fraction wasconcentrated by dialysisat00 against Aquacide

11-A(Calbiochem).

GlycerolGradient Centrifugation. Forestimationof

mo-lecularweight,theconcentrated phosphocelluloseenzyme was

diluted3-foldwith0.1 MpotassiumphosphateatpH 8.0,and layered on a linear 10-30% (vol/vol) glycerol gradient

con-taining0.1 MpotassiumphosphateatpH8.0, 2 mM DTT,and 0.02%Triton X-100. Centrifugationwasat48,000rpmfor12

hrat 10 inaSpinco SW-50.1 rotor. Fractionswerecollected

fromthebottomofthegradient andassayedforreverse

tran-scriptaseactivitywitholigo(dG)-poly(rC)astemplate. Bovine

serumalbuminservedas asedimentationmarkerinaparallel

gradient.

DNAPolymerase Assays. Assay mixturesfor polymerase activitywithsynthetic polymertemplatescontained (in100,l): 5 umol Tris-HCl atpH 8.0, 0.5,umolof MgCl2,0.1 ,umol of

DTT,andthefollowingcombinationsof polymerand dNTPs: 0.4,ugofoligo(dG).poly(rC)oroligo(dG)-poly(rCm),0.02,umol ofdCTPand1.0nmolof[3H]dGTP(4000 cpm/pmol);0.4

,ug

ofoligo(dT)-poly(rA)oroligo(dT)-poly(dA),0.02,umolofdATP

and1.0nmol of[3H]dTTP(4000 cpm/pmol).Inreactionswith

oligo(dG)-poly(rCm), MnCl2

(0.02,4mol)

replacedMgCl2. Assays with avian myeloblastosis virus (AMV-RNA)

con-tained (in 100,l): 5 Mumol Tris-HCIat pH 8.0, 0.8 4molof

MgCI2, 0.1 ,umol of DTT, 10 Agof actinomycin D (Sigma

Corp.),5jigofdistamycinA

(Calbiochem),

2;gofAMV 70S RNA,0.1

,4g

of

oligo(dT)12

18,0.1

,4mol

eachof

dATP,

dGTP,

dTTP,and5nmolof

[3H]dCTP

(1.5X

104

cpm/pmol).

Allreactions wereincubatedat360 for15-30minandwere terminatedbytheaddition of0.5 ml ofcold0.067 Msodium pyrophosphate/l MsodiumphosphateatpH 7.2,andthen0.5

ml ofcold80%trichloroacetic acid. Acid-insoluble

radioactivity

wascollected onmembrane filters and measuredin a scintil-lationcounter.

Terminal deoxynucleotidyl transferaseactivity wasmeasured

by the polymerization of[3H]dGTP in the absence ofa

com-plementarypolymer template. Reactionswerecarried out as

describedaboveexceptthat polymer dNTP combinationwas

replaced with0.4 ggofoligo(dG)lo18 plus0.02,mol ofdCTP and 1.0nmol of [3H]dGTP.

All syntheticoligo- and polynucleotides wereobtainedfrom Collaborative Research, Inc. and tritiated dGTP, dTTP, and dCTPwereobtained from NewEngland Nuclear. AMV-70S RNA was isolated from the purified virus as previously de-scribed (11).

Hybridization Reactions. Procedures for thehybridization reactionsand their analysis withSinucleasehave beenreported elsewhere (11). Conditions for Cs2SO4 equilibrium density

centrifugation (1) were modified by the addition of 0.02%

so-dium N-lauroylsarcosinateand

20,ug

each of E. coli DNAand

RNAtothegradients.

RESULTS

The data described below are based on the results of

indepen-dent enzymeisolations from four differenttumorcollections (labeled AthroughD). The fractionationandthepropertiesof the breast tumor polymerase didnot varysignificantly from

one preparation toanotherand hence the resultson eachone of the four need not be reported in complete detail.

Polymerase preparation A was isolated from a metastatic lesion in the liver of a patient with breast cancer. Nine grams of tumor were homogenized and the particulate material, bandingat adensity of 1.16-1.19 g/ml, wascollected as

de-scribed in Materialsand Methods. The recovered pellet was solubilized by the addition of Triton X-100 and then

fraction-ated throughapolyacrylamide/agarose gel (Ultrogel ACA-44)

column. Each column fraction was assayed for (i) DNA poly-merase with

oligo(dG)12

i8-poly(rC)

as a template and (ii) for terminal transferase, by using

oligo(dG)12-18

as a primer. Fig. 1A shows the fraction profile of these activities and their posi-tionsrelative to bovine serumalbuminandovalbuminwhich

were prerun on the same column as molecular weightmarkers.

Three peaks (two major and one minor) of DNA polymerase activity areobservedand it is evident thatthe first major peak also contains terminal transferase. The two major peaks are

found (Table1) to contain90%oftheappliedpolymerase ac-tivity and 3% of the protein as measured by thefluorescamine

procedure (12). To examine the reality of the separation of the polymerase activities observed in Fig. 1A,the two major peaks were pooled, dialyzed to reduce the phosphate buffer con-centration, and then chromatographed through a phospho-cellulose column with a linear (0.01-0.5 M) phosphate gradient with results as shown in Fig. 1B. We see that the polymerase activities again are resolved into two major peaks, one eluting at 0.08 M phosphate and the other at 0.18 M. It will be noted that again the second major polymerase peak is devoid of ter-minal transferase activity. The latter splits into two peaks, one associated with the first DNA polymerase activity at 0.08 M

phosphate, and another eluting by itself at 0.23 M phos-phate.

2.5 x s CL 1.5 0.5 20 40 60 80 FRACTION NUMBER 4" x 100 0.5 0.4 0.3

+',

0.2 X Cn 0.1 0 0.0 20 40 60 80 100 FRACTION NUMBER

FIG. 1. Purificationofthe breasttumorpolymerase. (A) Gel

fil-tration.Breasttumor(sample A)wasfractionedforparticulate

ma-terial ofdensity1.16-1.18g/cm3,This samplewassolubilizedand

filtered throughapolyacrylamide/agarose columnas describedin

MaterialsandMethods. Aliquots of each column fraction (25Al)were

assayedforpolymeraseactivity witholigo(dG)-poly(rC) (0---0) and

forterminal transferase activity witholigo(dG)(o-o).(BSA,

bo-vine-serumalbumin.)(B)Phosphocellulose chromatographyfractions

containingpolymerase activity from theUltrogelcolumnwerepooled,

dialyzed, and chromatographedonaphosphocellulosecolumn

(Ma-terials andMethods). The column fractions were assayed as in

(A).

Thesecond (0.18Mphosphate)

peak

of

polymerase

activity

observedonthephosphocellulosecolumnisnotfoundinnormal tissues(three samplesof breasttissue

and,

three samples from

spleens)orin benignfibroadenomasofthebreast (three pools

ofthree tofour.fibroadenomaseach) andappearstobe the

DNApolymeraseuniquetobreastcancertissue.The fractions

composing peak2of thephosphocellulosecolumnarefound

tocontain65%of theappliedDNApolymeraseactivityand8%

oftheprotein. Thesefractionsarepooled and concentratedas

describedinMaterialsandMethods toyield the breasttumor

Table 1. Purification of DNA polymerase froma

human breasttumor

Total Total Specific

Stage of protein activity activity

purification (mg) (pmol) (pmol/mg)

Viraldensity region 8.3 168 20

Polyacrylamide agarose 0.25 148 5.6x 102

Phosphocellulose 0.02 96 4.8x103

PdlymerasewaspurifiedfromtumorsampleAasdetailed in the

text.Allreactions containedoligo(dG)-poly(rC)astemplate andwere

incubatedfor30Ininat360. Totalactivitywascalculatedaspmolof

[3H]dCMPpolymerized.

Table 2. Synthetic polynucleotidesastemplates for the

breasttumorpolymerase

pmol[3H ] dNMPpolymerized [3H]- Prep. Prep. Prep. Prep.

Template

dNTP

A B C D Oligo(dT).poly(rA) dTTP 0.472 0.445 0.221 0.495 Oligo(dT)*poly(dA) dTTP 0.044 0.021 0.062 0.071 Oligo(dG)-poly(rC) dGTP 0.534 0.444 0.180 0.532 Oligo(dG).poly(rCm) dGTP 0.502 0.410 N.T. N.T. Oligo(dG) dGTP 0.010 0.008 N.T. N.T.

Polymerase reactionswereperformed with the indicated template

asdescribed inMaterialsandMethods. Incubationwasat360 for15

min'(prep.C)or30min(preps.A,B,andD). Polymeraseprep.Awas

isolated fromasingletumor(9g), B fromapool of threetumors(14

g),Cfromapool offourtumors(32 g), andDfromapoolof five

tu-mors(61g).N.T.,nottested.

polymerase for further characterization. Table 1summarizes

theyieldsofactivityandproteinateachofthe threesteps in

atypical-purification.

EvidenceThattheBreastCancerPolymerase IsaReverse

Transcriptase.Thereareseveral usefulcriteriawhich

distin-guishthereversetranscriptasesof theRNAtumorvirusesfrom

normal mammalian DNA polymerases. The viral reverse

transcriptases(13)showapreference foroligo(dT)-poly(rA)over

oligo(dT)-poly(dA) and they alsoacceptoligo(dG).poly(rC)and oligo(dG)-poly(rCm)asexcellent templatesfor thesynthesisof

poly(dG). Another,andmoreusefulcharacteristic, istheability

ofareversetranscriptase touseaheteropolymericRNAto

di-rectthesynthesisofafaithfulcomplementaryDNAas

dem-onstratedbyproperhybridizationof the DNAproducttothe templateusedinthe synthesis.

Theresponsesoffour of the breastcancerpolymerasestothe synthetic polyribonucleotidesaresummarizedinTable 2. The

resultsshow-apatternofactivitiescompletelyconsistent with thatobtained withreversetranscriptases isolatedfrom authentic

animalRNAtumorviruses (13). Thus,in allcases, oligo(dT)-poly(rA)issuperiortooligo(dT)-poly(dA)forthesynthesisof

poly(dT). Further, both oligo(dG)-poly(rC) and oligo(dG)-poly(rCm) were excellent templates for the formation of

poly(dG).Itshould- benoted thatinadditiontothe fourbreast

tumorenzymepreparationsdescribedhere,manyothers(more

than50)obtained from additionalpatientsinanongoingeffort

havebeen examined in thesame way atdifferent stages of

puritybyusingvariousmethodsof tissuefractionation, andthey

allexhibited theresponsepatterndescribedinTable2.

Theoperational definitionofareversetranscriptase requires

thedemonstration thatitcan use aheteropolymeric RNA to make aDDNA transcript. Theresponse(Table 3)of the breast

cancer

polymerase

totheRNA of theavianmyeloblastosisvirus

(AMV)isthat expectedfrom thesynthesisofaheteropolymeric

DNA. Leavingoutanyone,orall,oftherequiredthree

unla-beled deoxyribosidetriphosphates leads to the same virtual

disappearance ofsynthetic activityas occurs onomissionof the

RNAtemplate.

Themosttellingtestofaputativereversetranscriptase

re-actioncomesfromanexamination of thefidelityof the DNA transcript.Thisrequiresisolationof the[3H]DNA productand

challengingit inannealingreactions withthe RNA template usedinthe synthesis. A2-mlreaction,leadingtothesynthesis

ofapproximately3ng[3H]DNAat2 X107 cpm/,ug,wasrun

for 15 min with the DNA polymerase preparation A and

AMV-RNAasthetemplate.The[3H]DNA productwas

puri-l-A A~~~ ac

I.

n I % I

AI

1%

Table 3. Deoxynucleoside triphosphate and RNA requirements ofthe breasttumor DNApolymerase"

pmol

[3H]dCMP

polymerized Reaction A B Complete 0.200 0.170 -dATP 0.013 0.011 -dGTP 0.025 0.022 -dTTP 0.012 0.005 -dATP,dGTP, dTTP 0.008 0.011 -RNA 0.026 0.016

Polymerasepreparations AandBwereassayedforRNA-dependent

DNA synthesiswith AMV-RNAasthetemplate.The reactionswere

incubated at360for 30 minunder theconditionsgiveninMaterials

andMethods.

fied and recoveredbythe usualprocedures

(11)

and then an-nealed with AMV and Rauscher leukemiavirus70SRNAs.The

reactionproductwasexaminedbyseparationinCs2SO4 gra-dients andbyresistancetoSInuclease. Both methodsyielded less than 5% of the total product

annealing

totheunrelated

Rauscher leukemiavirus-RNA and between80and85% hy-bridizingtoAMV-RNA, thetemplateusedtodirect the syn-thesis. These resultssuggest, therefore,that the[3H]DNAisa

single-stranded complement of the AMV-RNA. A more in-formativeexaminationof the

[3H]DNA product

is

provided

by akineticstudyof theannealingreaction.Fig.2comparesthe

annealingkinetics oftwo

[3H]DNA products

toAMV-RNA;one

productwassynthesizedunder the direction ofAMV-RNAby AMV reversetranscriptaseand the other

synthesized

by

the human breast cancerpolymerase instructed

by

thesame

tem-plate. It isevidentthat the kinetics ofannealingto AMV-RNA

of thetwo DNAproductsareindistinguishable.

OtherProperties oftheBreast CancerPolymerase.

Vari-ationsintemperature and pHwereexaminedfor theireffects

on the activities of several preparations of the breast cancer

100 c 0 N 'a .0 >40 10-4 1-o3 10-2 10-1 10 Crt

FIG. 2. Hybridization kineticsofbreastpolymerase AMV-cDNA.

[3HJDNAsynthesized from AMV-70S RNAwith the human

poly-merase(0)orwithAMV DNA polymerase(0)wasannealed with

excessAMV-RNAtoCrt of0.8mol ofnucleotide-liter- per secat50%

in0.60Msodium chloride,0.06MsodiumcitrateatpH 6.5,0.5% so-diumdodecyl sulfate,and 42%formamide. Eachcurveisacomposite

oftwoannealing reactions of RNA concentrationsof 7.3 X10-:3and 1.8X101,ug/ml.Attheindicated Crt [initialconcentration of total RNA(molesofnucleotide/liter) Xtime(seconds)] values, aliquots (740 cpm)wereanalyzedwith-S,nuclease.Intheabsenceofadded

RNA,thebackground annealingofthehuman and AMV polymerase

cDNAswas10%and 4%,respectively.

x E

5 10 15 20

mmol/100gI

FIG. 3. Titration ofdivalent metalion.The activity ofthebreast

tumorpolymerasewasmeasuredasafunction oftheMg++(00) or Mn++ (---@) concentration with oligo(dT)-poly(rA) as the

template.These assays wereperformed withenzymepreparation B.

polymerase using oligo(dG)-poly(rC) andoligo(dT).poly(rA)

astemplates. The maximalrateofpolymerization occurredat 370 in 5 mMMgCI2over apH rangeof 7.0-8.5.

The divalentionrequirements ofreversetranscriptasesare

of some interestsincetheyservetodivide the virusesinto dif-ferent groups.Thus, allsix strainsofMMTVfrom avariety of

sources contain a reversetranscriptaseshowingastrong

pref-erenceforMg++ascompared withMn++ (14). Thesameholds

truefor the MPMV, thebromodeoxyuridine-induced guinea

pig virus (15), and thebovine leukemia virus(16).Incontrast, thereversetranscriptases of the murine leukemia andsarcoma

viruses(17, 18), the simian associated virus (18) and the hamster leukemia virus (19) function much more effectively in the presenceof Mn++.The cation requirement isexamined for the human breast cancer enzyme in Fig. 3 fromwhich it is evident that

Mn++,

at its optimum concentration, yields onlyabout one-seventh oftheactivity attainable with

Mg++.

Thus, the human breast cancerenzyme shows adivalent metalion re-quirement resembling the MMTV/MPMV virus group de-scribed above.

The molecular sizeof thebreastcancer enzyme was esti-mated by sedimentation through a linear (10-30% vol/vol) glycerolgradient. Theenzyme waslocatedby assayingfractions with AMV-RNA astemplate withresultsasdescribedinFig.

4. Theenzyme activitysediments between5and6S, slightly faster than thebovineserumalbuminmarker,therebygiving

amolecularweight estimate of about 70,000. Thisresult is also consistent with the relative elution positions of these same proteins onUltrogel (Fig. IA). Anumber ofenzyme

prepara-tions from different breast tumor sources yielded identical sedimentationvalues.

DISCUSSION

Theimmediate goal of the present investigation was to provide furtherinformationontheparticles detected (1, 2) in human breastmalignancies. These particlespossess many of the fea-turescharacteristic ofthe RNA tumor viruses including the following: (i)they sedimentatabout 600S and have a density ofaround 1.16g/ml;(fi) removal of an outer membrane with detergent converts them to"coreparticulates" with densities of 1.26g/mlorgreater; (M)the "cores" contain a 70S RNA that

candissociateinto35Ssubunits;(iv) the RNA of these particles-shares some sequence homology with the RNA found in MMTV,theetiologicagent of murinemammary tumors and with theRNAinMPMV, the virusisolatedfrom a spontaneous

x8 E C., 4 5 10 I5 20 25 FRACTION NUMBER

FIG. 4. Glycerol gradient centrifugation of the breast tumor

polymerase. Theenzymeactivity purifiedfromtumorsampleBwas

sedimentedona10-30%glycerol gradientMaterials and Methods.

A25 aliquot of eachgradient fractionwasassayed for polymerase

activitywitholigo(dG)-poly(rC)asthetemplate.Thearrowmarks

thesedimentation position of bovine-serum albumin (BSA) in a

parallel gradient.Similar resultswereobtained withenzyme

prepa-ration A.

breasttumorofarhesusmonkey; and finally,(v)the particles

containa reversetranscriptasewhichusesRNAas atemplate

tofabricateacomplementaryDNAcopy.

ThedatadescribedhereontheDNApolymerase purified

fromthe humancancerbreastparticles addfurther evidence

supportingsimilaritiestotheviralagentsassociated withsimilar malignanciesinthemouseandmonkey models.Theproperties

of the purified polymeraseare incomplete accordwith the publishedinformationonthevariousviralreverse

transcrip-tases.Thus, like the animal viralenzymes,thepurified human

breastcancerDNApolymerasepossessesthe followingfeatures:

(i)ithasasedimentation constantbetween5and6S,

corre-spondingtoanapproximatemolecularweight of70,000; (ii)

itsynthesizespoly(dT) about10timesbetterinthepresenceof

oligo(dT).poly(rA)ascomparedwitholigo(dT)-poly(dA); (iii)

both oligo(dG)-poly(rC) and, more significantly,

oligo(dG)-poly(rCm)areexcellenttemplatesfor thesynthesisofpoly(dG);

(iv) the humanenzymecanacceptaviralRNA(AMV-RNA) as atemplatetosynthesizea[3H]DNA productwhichbythe

hybridizationtestisasfaithfulacomplementarycopyas can

beobtainedbyusingthehomologousavianreverse

transcrip-tase;(v)the superiority ofMg++toMn++forthedetection of human polymerase activity mimics that observed with the

enzymesof theMMTVandMPMVgroup.

The data described herestemmed from theexaminationof fourenzymepreparationspurifiedtothesamestage-onefrom

ametastatictumorofthe liverand three frompoolsof breast

tumors.Allyieldedenzymeactivitieswith thesame

biochem-ical and biophysbiochem-ical properties. We have examined normal -tissues(5) and three pools (each containing between five and seventumors)ofthebenignfibroadenomas by the same frac-tionation techniques described here. These nonmalignant samplesyielded no enzyme activity with the properties of the reverse transcriptase purified from malignant tissue at any stage ofpurification.

The availability of the breast cancer enzyme in a compara-tivelypure statewill permit us to further delineate the bio-chemical features that relate it to, and differentiate it from, the reverse transcriptases of animal viral origins. Because of its implications for potential clinical usefulness, it would be of immediateinterest tofocus onthe immunological properties of the human enzyme withparticularreference to possible cross reactivitieswith any of the known animal reverse transcrip-tases.

Wewould liketoexpress our appreciation to Ms.I.Donevafor her

excellent technical assistance. This investigation wassupported by

GrantCA-02332andContractN01-6-1010awarded by the National

Cancer Institute.

1. Axel,R.,Schlom,J. & Spiegelman,S. (1972) Nature 235, 32-36.

2. Vaidya,A. B.,Black,M. M., Dion, A. S. &Moore, D. H. (1974) Nature249,565-567.

3. Colcher,D.,Spiegelman,S. &Schlom,J.(1974)Proc.Natl.Acad. Sci. USA 71,4975-4979.

4. Chopra, H.C. & Mason, M. M.(1970)CancerRes. 30, 2081-2086.

5. Schlom,J.&Spiegelman,S.(1971)Science 174, 840-843.

6. Axel,R.,Gulati,S.C. &Spiegelman,S.(1972)Proc.Nati.Acad.

Sci. USA 69,3133-3137.

7. Michalides,R.,Spiegelman,S. &Schlom,J.(1975)Cancer Res.

35,1003-1008.

8. Gerard,G. F., Lowenstein, P.,Green, M. & Rottman,F.(1975)

Nature256,140-143.

9. Witkin,S.,Ohno,T. &Spiegelman,S.(1975)Proc.Natl. Acad. Sci.USA 72,4133-4136.

10. Trautschold, I., Werle, E. & Zickgraf-Rudel, G. (1966)

Arzneim.-Forschung. 16, 1507-1515.

11. Weiss, M. J., Sweet, R.W.,Gulati,S. C. &Harter, D. H.(1976)

Virology71,395-401.

12. Udenfriend,S.,Stein, S.B6hlen,P., Dairman, W.,Leimgruber, W. &Weigele,M.(1972)Science178,871-872.

13. Gerard,G.F., Rottman,F. &Green,M.(1974)Biochemistry13,

1632-1641.

14. Marcus, S. L., Sarkar,H. & Modak,M.J. (1976)Virology 71,

242-254.

15. Schlom,J., Colcher,D.,Drohan,W.,Kimball,P.,Michalides,R. &Schochetman,G.(1976)inBreastCancer:TrendsinResearch

andTreatment, eds. Henson, J.C.,Mathheiem,W. H. &

Ro-sensweig,M.(RavenPress,NewYork),pp.11-51.

16. Kettmann, R., Portetelle, D., Mammerickx, M., Cleuter, Y.,

Dekegel,D.,Galoux,M.,Ghysdael,J., Burny,A. &Chantrenne,

H.(1976)Proc.Natl.Acad.Sci. USA73,1014-1018.

17. Moelling,K.(1974)Virology62,46-59.

18. Abrell,J.W. &Gallo,R.C.(1973)J. Virol. 12,431-439. 19. Verma,I.M.,Menth,N. L.,Fan,H. &Baltimore,D.(1974)J.