Different forms of simian virus 40 large tumor antigen varying in their affinities for DNA.

JOURNAL OF VIROLOGY,May1982,p.456-466 Vol.42, No. 2 0022-538X/82/050456-11$02.00/0

Different

Forms

of Simian Virus 40 Large Tumor Antigen

Varying

in

Their Affinities for DNA

DAVIDGIDONI,1 ARNOSCHELLER,2 BETHBARNET,2 PETROS HANTZOPOULOS,2 MOSHE

OREN,1ANDCAROLPRIVES2*

Virology Department, Weizmann InstituteofScience, Rehovot, Israel,1 and Department of Biological Sciences, ColumbiaUniversity, New York, New York100272

Received19October1981/Accepted14January1982

Invariouspermissive monkeycell lines infected with simian virus 40thereare

two major forms oflarge T antigen which differ in their rate of sedimentation

through sucrose gradients. The lighter (5 to 7S) form sedimented slightly more

rapidly than the 4S tRNA marker, whereas the heavier (16S) form sedimented slightly more slowlythan the 18S rRNAmarker.Thesmalltantigen did not form complexes whichsedimented asrapidlyasthose formedbythe largeTantigen.

The 16S Tantigenformwasconvertedto theslowly sedimenting5 to7S form in

thepresenceof1.0 M NaCl. The majorityoflarge Tantigen synthesized in cell-freeprotein-synthesizing systems primed by mRNA isolated frominfected cells

sedimented as the 5 to 7S form even when premixed with excessquantities of cellularTantigen. The formation of the16Sform ininfected cells did notrequire ongoingviral orcellular DNAreplication because considerablequantities of this T antigen class were producedin thepresenceof DNAsynthesis inhibitors,such as

cytosinearabinoside. Both 5 to7Sand16Sformscould beisolatedseparately and, therefore, each could beanalyzed as to its individual properties. The 5 to7S T

antigen form boundmore efficientlyandtightlytoDNAandhadspecific affinity forsequences at the viraloriginofreplication, whereasthe16Sformbound less efficientlyto DNAandexhibited verylittlespecificity fororigin-containing DNA sequences.It isthereforelikelythat theactiveDNA-bindingspecies of T antigen isolated from infected cells is the5 to 7Sform.

The early region of the genome of the DNA tumorvirus simian virus 40(SV40) encodes two

polypeptides, the large T and small t antigens

(forareview see reference 42). Genetic studies have shown that large T antigen is required for initiating viral DNA replication (36) and viral L-strand-specific RNA transcription (4) and for initiation (12) and maintenance (2, 16, 25, 37) of the transformed phenotype in nonpermissive cells. Inaddition, this geneproduct exerts nega-tivefeedbackcontrol over early mRNA synthe-sis (31, 39). Small t antigen has noapparent role inthelytic cycle of thevirus but appears to play

a role, by slowing tumor growth, in virus-in-duced malignant transformation (17). Although little is known about the properties of small t antigen, considerable information has been ac-quired regarding the biochemistry of the large T antigen. Studies utilizing either authentic T anti-gen fromlytically infected or transformed

cells,

ortheclosely related D-2 protein, have shown it

tobe aphosphoprotein (38) exhibiting both tight

(specific) (23, 34, 40) and loose (nonspecific)

affinities for DNA (3, 23), as well as having ATPaseactivity (41). In addition, both earlier(3,

13, 15, 24, 27) and more recent (5, 6, 18, 28)

papers have described theability of theT anti-gen protein to form rapidly sedimenting com-plexes. Asbothbiological andbiochemical evi-dence supports the idea that T antigen is a multifunctional protein, we sought to isolate different forms of the Tantigen which may have different functions. Morespecifically, we isolat-ed and characterizisolat-ed twoformsof the Tantigen

frominfected cells which differ in their sedimen-tation properties and affinities for DNA.

MATERIALSANDMETHODS

Cells and viruses. BSC-1 or CV-1 lines of African greenmonkeykidney cellswereused,both for infect-edcellextractsandtogrow SV40 virusstrains 777 and 776.

Labeling and extractions of cells. Extraction and labeling of infected cells were done by previously publishedprocedures (29). Generally, cultures of4x 106cells werelabeled with[35S]methionine (100FiCi/

ml) in methionine-free medium from 42to46or 44 to 46 hpostinfection with SV40. All subsequent opera-tionswereperformedat4°C. Cellswerewashed three times inphosphate-buffered saline and then lysed in 0.6 mlof HIPbuffer(pH 8.2)containing0.14 MNaCl, 0.02 M N-2-hydroxyethylpiperazine-N'-2-ethanesul-fonic acid(HEPES),0.001MMgCI2, and0.5%

Noni-456

on November 10, 2019 by guest

http://jvi.asm.org/

DNA BINDING OF SV40 T ANTIGEN OLIGOMERS

det P-40. Phenylmethylsulfonyl fluoride and 1-1-to-sylamide-2-phenylethyl chloromethyl ketone (TPCK) wereadded to a final concentration of 0.25mg/mlof extract immediately before lysis. Nuclei and debris were removed by centrifugation at 2,000 rpm for 2 min.

Sucrosegradientsedhnentation and immunoprecipi-tation. Cytoplasmic extracts (200 p.1) were layered onto5-mlgradients of 5 to20%sucrose in HIP buffer containing 50 p.g offreshly added TPCK per mland 100,ug ofphenylmethylsulfonylfluoride perml. Gradi-entswerecentrifugedat4°Cfor 180minat48,000 rpm in a BeckmanSW50.1swinging bucket rotor. Ten 520-,ul ortwenty260-p,lfractions werecollectedper gradi-ent. After removal of15-p.l samples of eachgradient fraction for direct analysis by polyacrylamide gel electrophoresis (PAGE), the remaining portion was immunoprecipitated with 10,ulofanti-SV40Tserum for1h at 4°C.Immune complexeswereprecipitated by theaddition of 50

IL1

of a10%suspension of formalde-hyde-fixed Staphylococcus A bacteria for 15 min. Immunoprecipitates were washed, suspended in elec-trophoresissample buffer, and subjected to PAGE as previously described (29). Gels were fluorographed, dried, and exposed to Kodak XR-5 film.

mRNA preparation and translation. Cytoplasmic polyadenylic acid-containing RNA was purified from SV40-infected cellsandtranslatedin thenuclease-free reticulocyte lysate system according to previously publishedprocedures (30). mRNA (20 ,ug) was added per 500Illof translation reaction mixturecontaining 400

IuCi

of[35S]methionine.Translation products were layeredonto 5 to20%o sucrose gradients, centrifuged throughsucrose gradients,andimmunoprecipitatedas above.

DNA-bindingexperimepts. Binding of T antigen in sucrose gradient fractionsto DNA-cellulosecolumns wascarried out byprocedures described by Oren et al. (23). Samples (200 ,ul) of[35S]methionine-labeled

infected cell extracts were centrifuged through five sucrosegradients run in parallel, separately fractionat-ed, and stored. One set of gradient fractions was immunoprecipitated with anti-T serum, subjected to PAGE, and then autoradiographed to identify peak tubes of T-antigen-sedimenting forms, which were thenpooledseparately from the other individual gradi-ents. Samplestobebound to DNA wereadjustedto pH 7.0 with2-1lincrementsof 0.1 N acetic acid and loaded onto a 1-ml calfthymus DNA-cellulose column orF-161SV40multioriginvariantDNA-cellulose col-umn(23)which had beenpreequilibratedwith 10 mM potassium phosphate-0.1MNaCI-109o(vol/vol) glyc-erine-0.5% NonidetP-40, pH 7.0. Afterallowing 30 minforthesampletobind,thenonbound fractionwas

collected,and the columnwaswashedwith

equilibra-tion buffer.Boundproteinswereelutedstepwisewith bufferscontaining0.4 Mand 1.0 MNaCI.Nonbound and boundelutedfractionswerethenadjustedto0.15 MNaCI, immunoprecipitated with anti-T serum, and subjectedtosodiumdodecyl sulfate-PAGEand

auto-radiographyasabove.

DNAbindinginsolutionwasperformed byrecently

describedprocedures (A. Scheller,L.Covey-Nichols,

B.Barnet, C. Prives, submittedforpublication).

Su-crosegradientfractionswereadjustedtopH6.5 with

2-01

samples of0.1 Nacetic acid. Five nanograms of

BstNl-digested SV40DNAlabeledwith 32Pby

nick-translation (32), as well as 5 ,ug ofsheared salmon sperm DNA, was added to each fraction for 1 h followed by immunoprecipitation of the T antigen DNA complexes with 10 pl of anti-T serum and StaphylococcusAbacteria as above. After the com-plexes had been washed three times, DNAfragments were released from the immune complexes with 50Il1

of buffer containing 0.01 M Tris, 0.001 M EDTA, and 0.5% sodium dodecyl sulfate (pH 7.4), subjected to electrophoresis through 2.0%'o agarose gels in Tris-boratebuffer,thendried,andautoradiographed.

Materials. [35S]methionineand

[32P]deoxynucleoside

triphosphateswerepurchasedfromtheRadiochemical Centre, Amersham, England. TPCK, phenylmethyl-sulfonyl fluoride, hydroxyurea, 2'-deoxy-2'-azidocyti-dine, aphidicolin, and cytosine arabinoside were boughtfrom SigmaChemical Co., St. Louis, Mo., and BstNl restriction endonuclease was purchased from New EnglandBiolabs, Beverly, Mass. Staphylococ-cus Abacteria were purchased from Calbiochem, La Jolla, Calif., as Pansorbin. Anti-SV40 T serum was obtained from the Research Resources Branch, Na-tional Cancer Institute, NaNa-tional Institutes ofHealth, Bethesda,Md.

RESULTS

Existenceof Tantigen as two

major

sediment-ing forms in infected monkey cells. To extend

studies describing different oligomericforms of

T antigen, extracts of

[35S]methionine-labeled

BSC-1 cells were subjected to sedimentation through sucrose gradients. Individual

gradient

fractions

werecollectedand

immunoprecipitat-ed with anti-T serum, followed by PAGE and

autoradiography. Undertheconditionsof

infec-tion,labeling,

extpWption,

andcentrifugation

de-scribed herein, tpe

large

T antigen appears as

two major sedimenting classes (Fig. 1). The

lighter class cosedjienting with 5S RNA and slightly less

rapidly'

than aldolase (molecular weight, 159,000), wastermedthe 5 to7S form.

The heavier class,

se4imenting

lessrapidlythan

itS

rRNA or thyroglobulin

(molecular

weight,

670,000),

was termed the 16Sform. Itwas not

feasibletoassign definitive sedimentation coeffi-cients or molecular weights for these

forms,

isolated as they are from the crude extracts of infected cells. In many

experiments

a

45,000-molecular-weight

polypeptide

was immunopre-cipitated from all

gradient

fractions. This may

well

bethemajorcapsid

protein

VP-1which has this molecular weight, reactswith many anti-T

sera(35),andforms

rapidly

sedimenting

assem-blycomplexes

(26).

The small t

antigen,

although

sedimenting

somewhat more

rapidly

than

expected

for a

17,000-molecular-weight

polypeptide,

was not

detectable in

gradient

fractions

containing

the 16SformoflargeT

antigen

andtherefore does

not

participate

in the

rapidly

sedimenting large

T

antigen

complexes.

The sedimentation

proper-ties oflarge Tand smallt

antigens

were

repro-VOL.42,1982 457

on November 10, 2019 by guest

http://jvi.asm.org/

J. VIROL. 458 GIDONI ET AL.

18$

4, :l

a

_S. - s1..

b 3s r

" _ ,,., - ...

'S,8 ,"110 4..

tg :pI :._". #W. .:

Ju. 1. Sucrosegradient sedimentationofSV40T

antigen. [35S]methionine-labeled cell extracts were

subjected tocentrifugation through5to20%sucrose

gradients asdescribed in the text. Twenty fractions

were collected, and 10-,ul samples of each fraction

weredirectlyanalyzed by PAGE andautoradiography

(a).Theremainderof eachsamplewas immunoprecip-itated withanti-SV40Tserumandsimilarly analyzed

(b).Twoparallelsucrosegradientswererun, contain-ing [3H]uridine-labeled rRNA markers in one and purified preparations of aldolase andthyroglobulinin the other. Peak tubes of 5S, 18S, and 28S rRNA markers were nos. 4, 11, and 17, respectively, and

peaktubesofaldolase(molecular weight, 159,000) and thyroglobulin (molecular weight, 670,000)werenos.5 and11,respectively.

duciblyobserved in several SV40-infected

mon-key lines, including BSC-1, CV-1, TC-7, and

AGMK cells. The relative proportionof the 5to

7S and 16S large T antigen form varied among

differentmonkey cell lines.Generally, therewas

more [35S]methionine-labeled 5to7S form than

16S formimmunoprecipitatedfromsucrose

gra-dientfractions. However, therewas more

rapid-ly sedimentingthan slowly sedimentingT

anti-gen when this protein was measured by the

enzyme-linked immunosorbent assay,

per-formedasrecently

published

(11,

43),

indicating

that muchoftheaccumulatedT

antigen

existsas heavier complexes (data not shown). As the

experiments in this study measured the

[35S]methionine-labeled

T antigen, it should be kept in mind that all the data

presented

referto

that population ofT antigen molecules

synthe-sizedduring thelabelingperiod of2to4 hbefore extraction.

Tantigensynthesizedinvitro.Studiesinother laboratories (5, 6, 13), as wellas our own

(28),

have shown that tsA mutant Tantigendoesnot

form rapidly sedimenting complexes at the

re-strictivetemperature,

suggesting

that the gener-ation of the 16S form is related to a

biological

function ofT antigen. To further explore this idea, weexaminedthesedimentation

properties

ofT antigen

synthesized

in acell-free

protein-synthesizingsystemprimedwithSV40E-strand

mRNA.This Tantigenwouldserve as an

exam-ple of nonfunctional protein in that it has no opportunity to act in viral and cellular DNA

replication or RNA transcription. Moreover, it may notundergo theidentical

post-translational

modificationscharacteristicof Tantigen isolated

from infected cells.

When Tantigen synthesized inthe nuclease-treated reticulocyte

lysate

protein-synthesizing

systemwascentrifuged throughasucrose

gradi-ent, noevidenceoftherapidlysedimentingform was detected (Fig. 2). Even afterco-incubation

and co-sedimentation of the cell-free

reticulo-cyte extractwith unlabeled extracts of infected

or uninfected cells, the in vitro-synthesized T antigen existed only as the 5 to 7S form. The results shown inFig. 2, in which 10 fractions per

gradient were collected, were confirmedin

ex-periments with 20 fraction gradients, such as those normally used in otherexperiments.This supports the idea that the 16S form of T antigen isaproperty of native,functional protein. Either

a specific post-translational modification or some aspect of the biologicalfunction of the T antigen protein is responsible for theability of this viral protein to form rapidly sedimenting complexes.

16S

Tantigenformation.Neither tsAmutant T antigen at restrictive temperature nor in vitro-synthesized T antigen forms oligomers. Both are nonfunctioning T antigens, the former due to its conditional defect, the latterbecause of its lack ofopportunitytofunction or lack of proper post-translationalmodification orboth. Another link betweenthese twonon-oligomerizingTantigens isthe absence of DNAreplication. tsA mutant-infected cells failed to initiate viral or cellular

DNAreplication at thenonpermissive tempera-ture, andclearly there was no DNAreplication

on November 10, 2019 by guest

http://jvi.asm.org/

DNA BINDING OF SV40 T ANTIGEN OLIGOMERS 459

i

a

2 4 6 8 10

r

.. Ams Wa 0

T-.

ep3q.

C

.. .*a a

C

t 4 6 8 10

. .. 6

o..:.. !1

_ *+* a | j

i- :s.

e

:M.'{:BiE'-t

..::Mr

******

X .F

!

tv :

__....-

_ J

2 4 6 8

T- n,

[image:4.505.115.404.69.515.2]

d

FIG. 2. Sedimentation of T antigen synthesized in vitro. SV40Tantigenwassynthesized inthereticulocyte

translation system (0.5-ml reaction volume) in response to mRNA from SV40-infected cells. Translation products (100 gxl)weremixed with 100 Il of HIP buffer (a); 100 II of infected unlabeled cellextract(b); 100 ILl of uninfected unlabeled cellextract(C) for 1 hat4°C followed bysucrosegradient centrifugation,

immunoprecipi-tation, and analysisasdescribed in thelegendtoFig.1.[35S]methionine-labeled Tantigen (100 ILI)from SV40-infectedextractwassimilarly analyzedon sucrosegradient (d).Ten fractionswerecollectedfrom eachgradient. That 5 to 7S and 16S T antigen forms could be distinguished under these sedimentation and fractionation conditions is seen in (d). Peaktubes of5S, 18S, and 28S rRNA markers run in parallelwere 3, 6, and 9, respectively.

in reticulocyte lysates. This led us to question whetherornotthere is arequirement for

ongo-ing DNA replication in order for the oligomeric T antigen to form. To test this, extracts of [35S]methionine-labeled cellswereprepared 10h postinfection, at a time considered to precede

the onset of virus-induced cellular and viral DNA replication. After centrifugation ofthese earlyextractsand anti-Timmunoprecipitation of gradient fractions, we observed no detectable difference in the relative proportions of16Sand 5to7S forms when compared with thoseof the

10 VOL.42, 1982

on November 10, 2019 by guest

http://jvi.asm.org/

460 GIDONI ET AL.

I

I,

IIi %I, i*

b-o

Top

X

II

II

I It

'O

'I

tb~

IQ,IXe Bottom

10,s D0

zL0--2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2u

FRACTION NUMBER

FIG. 3. Effect ofcytosinearabinosideonTantigensedimentation. Cultures of BSC-1 cells to whichcytosine arabinoside(20Fg/ml)had been addedatthe time of infection withSV40werelabeled with[35S]methionine42to 44hpostinfection,extracted andcentrifugedthroughsucrosegradientsasdescribedin the text and in thelegend to Fig. 1. Anti-Timmunoprecipitates of 20gradientfractions weresubjectedto PAGE andautoradiography, using Kodak XR-5 film.Eachslot,representingtheimmunoprecipitateofagradientfraction,wasscannedwitha Gilforddensitometer,andareasundercurvesrepresenting largeTantigen gelbandswerecut out andweighed.In parallel cultures treatedwithcytosinearabinosideasin thisexperiment,DNAsynthesiswasinhibitedby95%. 5S and 18S rRNA marker peaktubeswere4 and10,respectively. Symbols:0,culture treated with cytosine arabinoside; 0,control untreated culture.

typical 44-h late extract. However, as thecells

werenotsynchronized and therefore theremay

havebeen someDNAreplication induced even

at10 hpostinfection, several inhibitors ofDNA synthesis were used to further ensure the ab-senceof viral and cellular DNAreplication. All compounds tested, including hydroxyurea,

2'-deoxy-2'-azidocytidine (1), aphidicolin, and cytosine arabinoside yieldednoreduction in the relativeamountof labeled 16S T antigen, under conditions where DNA synthesis was reduced

by 90to95%.Indeed, inmany casestherewas a

significant enhancement of the 16S form of T antigen. Aparticularly dramatic example of this increase isseenin Fig. 3. One possible

interpre-tationoftheseresults is that T antigen oligomer-ization may, in fact, be an alternative to its

functioningin DNA replication.

Sensitivity of 16S T antigen form to increased ionic strength. Toelucidate the properties ofthe bonds forming the oligomers of T antigen, we

tested different extraction and centrifugation procedures. Pulse-labeling and pulse-chase

ex-periments showed that the 5 to 7S form is

labeled beforethe16Sform(datanotshown),as

has been reported by others (5, 6). Therefore,

the relative proportions of [35S]methionine-la-beled T antigen 5to7S and16Sformschanged withthe duration of the labeling period, consis-tentwiththeobservation thatmostofthe

accu-mulated T antigen measured by the

enzyme-linked immunosorbent assay exists in rapidly sedimenting complexes. Under similarlabeling,

extraction, and centrifugation conditionsas

de-scribedforFig. 1, the relativeamountof the16S

form wasnot appreciably affected by the pres-ence or absence of0.5% Nonidet P-40, 5 mM MgCl2 or CaCl2, or 5 mM dithiothreitol or

,B-mercaptoethanol. Moreover,the amountofthe

16Sform wasnotdiminishedwhencellextracts

were incubated with quantities of micrococcal

nuclease sufficient to digest 40 ,ug of rRNA markers. Thisindicates that the T antigen in the

16Sform isnot associated with long, nuclease-sensitive stretchesof nucleic acid.

However,sometreatmentdidaffectthe integ-rity of the 16S form. When extractsof infected cells were adjusted to 1.0 M NaCl and

centri-20 19~ -18 x

E

o 17

.17

12

-i12

CZ

D 10

_cc.)_

u0

z

<

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.505.100.408.52.334.2]

DNA BINDING OF SV40 T ANTIGEN OLIGOMERS

fuged through sucrose gradients containing this salt concentration, the oligomeric forms of T antigen vanished (Fig. 4). Indeed, the 5 to7ST antigen form was also seen to sediment even moreslowly than normally observed, suggesting that this form may itself beheterogeneous, per-haps consisting of monomers and dimers, the latter beingsimilarlydissociated at 1.0 MNaCI.

Thus, T antigen oligomers are unstable at higher salt concentrations and must be generated through bonds sensitive to increase in ionic strength. The 45,000-molecular-weight protein tentativelyidentified as VP-1 is also apparently sensitive to the increased saltconcentrations.In

addition,whenextracts were adjusted to pH 6.0 andcentrifugedthrough sucrose gradients at this

pH,there was a considerablereduction but not

an entire disappearance of the amount of the

[35S]methionine-labeled

16S T

antigen

form

(data not shown). This pH effect was not ob-served above pH 6.5.

Stability

of 5 to 7S and16S T antigen forms in

vitro. Several lines of evidence point to the

possibility that SV40 T antigen is a

multifunc-tionalprotein whose various biological activities may reflect different domains or forms of the molecule. As the protein exhibits heterogeneity

with respect to its sedimentation properties, it was of interest to test whether these forms differed with respect to their biological and

2 4 6 8

go ww ,

-

T-biochemicalproperties. To assess any such

dif-ferences,assurance was sought that these forms were notinterconvertible in vitro. For example, as oligomerization of many proteins can be a

concentration-dependentevent, it might be sus-pected thatdilution of the 16S form may cause itsreconversiontothe 5 to 7S form. To test this

possibility,aswell as to test the stability of the 5 to7S form, extracts ofSV40-infected cellswere

subjected to sucrose gradientcentrifugation as

describedinFig. 1. The peak fractions of the 5 to 7S and 16Sformswerethendiluted twofold and threefold, respectively (to dilute the sucrose

concentrationtobelow5%),and each was sepa-rately subjected to a second centrifugation

through sucrose gradients(Fig. 5). Both5 to7S and 16S T antigen forms, even after dilution,

retained theiroriginalsedimentation properties;

5 to 7S T antigen remained as such, whereas virtually no 16S Tantigen was converted to the more slowly sedimentingforms. Nor did there appear to be any conversion ofthe 5 to 7S to

16S form in vitro: coincubation of isolated

[35S]methionine-labeled

5 to 7S peak fraction

with an unlabeled 16S peakfraction did not yield any labeled 5 to 7S T antigen forming heavier

complexes upon re-centrifugation (data not

shown).

DNA-bindingproperties of different

sediment-ing

forms

of T antigen. Many ofthe roles ofT

10 12 14 16 18 M

U

94 68

45

do 30

7 el-.-.

..1...;.i.

[image:6.505.135.392.401.598.2]

4_-12

FIG. 4. SedimentationofSV40Tantigenathigh ionic strength.Extracts of[35S]methionine-labeled infected cellswerepreparedasdescribed in thetext.Afteradjustingthe concentration ofNaClin the extract to 1.0 M, 200 p. wascentrifugedthrough5to20%sucrosegradientsin buffer containing 1.0 MNaCl,0.02 M HEPES buffer, and 0.001 MMgC12 (pH7.4).Nonidet P-40wasomitted fromgradientsolutions due to co-precipitation with 1.0 M salt in the20%osucrose-containingsolution. However, the absence of Nonidet P-40 from sucrose gradients containing0.15 MNaCl didnotmeasurablyaffect theproportionsof 5 to7Sand16ST antigen forms (see the text). Thepeaktubes of 5to7S and16S Tantigenformsruninparallelunder identical conditions as in Fig. 1 were 4and8,respectively.

VOL.42, 1982 461

on November 10, 2019 by guest

http://jvi.asm.org/

462 GIDONI ET AL.J.Vo.

&~~~~~~~~~~~~~~~~~~~~~~~~~~~...

b ~~~~~~~~~~~~~~~~...2.

+

~~DNA.

Two DNA-binding assays were

em-ployed,

both ofwhichtestfor

tight

binding ofT

antigen

toDNA. In thefirstassay,

specific

and

T

nonspecific

bindingwasexaminedbytestingthe

differential salt-sensitive bindingofT antigento

SV40

origin-containing

(SV40

multiorigin

vari-ant[MOVIand

origin-minus

(calfthymus)DNA linkedto cellulose (23). When thismethod was

developed,

itwas found that therearedifferent

classes ofT

antigen

whichexhibit

varying

affini-ty forDNA. To examine whether these classes

might

be related to different

sedimenting

forms

ofTantigen, the

peak

tubes ofthe 5 to 7S and

16S form werecollected after sucrose

gradient

sedimentation, and each bound to and eluted from calf

thymus

orSV40MOVDNA-cellulose columns

(Fig.

6). Ithas been

previously

shown that the amount of DNA onthe calfthymus or

MOV DNA-cellulosecolumns isinconsiderable.

excess ofthe bindable T

antigen

isolated from

comparable

quantities

of infectedcells (23, 29).

Under the binding (pH 7.0; 0.15 M

NaCI)

and elution (pH 7.0; 0.4 M and 1.0 MNaCl)

condi-tions used,itwasfound that the 5to7S formof

[image:7.505.60.245.61.338.2]

large T antigen binds calf thymus and MOV

FIG. 5. Re-sedimentation of5Sand 16S Tantigen

forms. Sucrose gradient sedimentation of

[35S]methi-onine-labeledinfectedcellextracts was performedas

described in thelegendtoFig.1.Peak tubesof5 to75 and16STantigenforms(generallytubesnos.4and10) were diluted twofold and threefold, respectively, in HIP buffer and subjected to sedimentation through

sucrose gradients as before. Two parallel sucrose gradientsof diluted 55and16Speak tubeswererun, andthe corresponding fractions of the twogradients werepooledandimmunoprecipitated. (a) Re-sedimen-tation of 5to7S Tantigenpeak tube; (b) re-sedimenta-tion of 16S T antigen peak tube. Inparallel sucrose gradientsedimentationofrRNAmarkers,55, 185,and 28SrRNApeaktubeswere nos.4, 11,and16,

respec-tively,asindicatedbyarrows.

ci O c d e f

;,.M..gm

.: ... :.

-M.:....

T...- .:

g h i 1kI

4 O

antigen

in the lytic cycle

imply

its interaction

with DNA,

specifically

with the viral origin of

replication. Theabilityof Tantigento specifical-ly bind DNA sequences at the viral origin in vitro is well documented through several ap-proaches (34, 41). Ithaspreviouslybeen shown that there are different subsets of T antigen molecules which vary in their affinity for DNA (8, 23, 29).The existence of Tantigenastwoor

more

sedimenting

forms raised the question as

towhether these forms may differ in their affini-ties for DNA and thus may be related to the DNA-binding heterogeneityof Tantigen.As the 5to75and 165 classes, when separately

isolat-ed,retained theirabilitytosedimentaslightand heavy forms, respectively (Fig. 5), itwas possi-ble to test their individual interactions with

FIG. 6. Binding of55 and165Tantigenformns to calfthymus and MOV DNA-cellulose.

-Five

sucrose gradientsof

200-pzl

samplesof

[35S]methionine-labeled

infectedcellextractswere runinparallel.PeakS5to7S

(tube no. 4) and 16S (tube no. 10) fractions were pooled and adjusted to pH 7.0, and each poolwas

dividedintotwoportions. DNAbindingof 5to7S(a through f0 and 16S (g through 1) T antigen to calf

thymusDNA-cellulose(a, b,c,g,h, i)orF-161MOV DNA(d,e,f,j, k, 1)wastestedby bindingand elution

pH7.0inbuffercontaining0.1 MNACI (a, d, g,j), followedbyelution inbuffercontaining0.4 MNaCl(b,

e,h,k),then inbuffercontaining1.0 MNaCl(c, f,i, 1).

Allelution fractions wereadjustedto0.1 MNaClby dilution, immunoprecipitated with anti-T serum, and

analyzedbyPAGE andautoradiography.

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:7.505.272.439.366.523.2]

DNA BINDING OF SV40T ANTIGEN OLIGOMERS 463

DNA, whereas small t antigen does not. The 5 to 7S T antigen, which bound to DNA, could be eluted from calf thymus DNA with 0.4 M

NaCI,

butrequired higher ionic strength (1.0 M NaCI) tobe eluted from MOV DNA, consistent with earlierobservations of whole cell extracts of

T-antigen-containing infected cells (23). In con-trast, the 16S form bound very inefficiently to

calfthymus DNA and even more poorly to MOV DNA, in thatconsiderably more of this T anti-genclass did not bind to the column at all when compared to the proportions of bound and non-bound 5 to 7S T antigen. In contrast to the 5 to

7Sform,nodetectable tight (1.0 M

NaCl-sensi-tive) binding of the 16S form to MOV DNA was

observed.Thus, these two forms clearly differ in their interaction with DNA immobilized on cel-lulose columns.

We also tested the binding of T antigen to

various DNA sequences in an assay in which extractsfrominfected cellscontainingTantigen areboundto32P-labeledrestriction fragments of viral SV40 DNA in solution (19; Scheller et al.,

submitted forpublication). The T antigen-DNA

fragment

complex is then immunoprecipitated

withanti-SV40 T serum(Schelleret al.,

submit-tedfor publication).Incubation of BstNl-digest-edSV40 DNA comprising 13 fragments,

includ-ing the 311-base-pair G fragment spanning the viral origin ofreplication, with extracts of

SV40-infectedcellsundersuitable binding conditions,

yieldedconsiderablequantities of the origin

se-quence-containing G fragment specifically bound to Tantigen.Totestwhether the different

sedimenting forms ofT antigen varied intheir

ability to specifically bind to origins, gradient

fractions were bound to 32P-labeled BstNl-di-gestedSV40

fragments

in solution(Fig. 7).

This experiment differed from the column-binding experiments in two major ways: first,

the kinetics of binding in solution were more rapidthanwhenDNAwasimmobilizedon

cellu-lose; second, whereas the column-binding

ex-periment

provided

information about labeled (newly

synthesized)

T

antigen

molecules, the

solution-binding experiment measured the

spe-cific-origin-binding activity

ofthe total

accumu-lated T antigen in each

gradient

fraction. The

resultsconfirmed and extendedthose obtained

with the DNA-cellulose binding assay.

Only

fractions in the 5 to 7S region of the gradient boundtoBstN1 fragment

G;

more

rapidly

sedi-menting forms exhibited little or no

specific

affinity for the

origin-containing fragment.

It is

noteworthythattubeno.4showed

considerably

greater Gfragment

binding

than did tubeno. 3.

Consistentwith theresults of the NaCl dissocia-tion experiment

(Fig. 3),

which

suggested

that the5 to7S classmaybeheterogeneous,it can be

suggestedthat the 5to7ST

antigen

may consist

of both monomers and dimers, of which the latter are the major origin-binding form.

DISCUSSION

Our experiments indicate that different forms

of T antigen vary in their affinity for DNA. Under the DNA-binding conditions that we used, the 5 to 7S T antigen form could bind to DNA with high affinity, whereas the 16S form could not. These data are somewhat at odds with those of Myers et al. (21), who provideevidence from electron microscopic analysis that prepara-tions ofpurified and concentrated D-2 protein

AM 2 4 6 8 10 12

I _

T 4..

lA IS 18 20

_

M2n5e , _

B M 2 3 4 5 -0 o 9C ! mv a 35 CO _M

A-

D- C--

F-FIG. 7. Bindingof varioussedimenting forms ofT antigento DNAfragments in solution. Twoparallel

sucrose gradients of

200-1.l

portions of

[35S]methio-nine-labeled and unlabeled infected cell extractswere run.Gradient fractionsof[35S]methionine-labeled ex-tractswerecollected andsubjectedto immunoprecipi-tation and analysisas described in thelegendtoFig.

1(A),orgradientfractions of unlabeledextract were incubated with 5 ng of32P-labeled BstNlSV40 restric-tionfragments, followed by immunoprecipitation of the Tantigen DNAcomplexas described in thetext (B). Tubes from fractions 6 andupwardsweredoubled in theDNA-bindingassay, asshown in thefigure,to

compensate for the greaterquantities of

[35S]methio-nine-labeled T antigen present in the more slowly

sedimenting fractions.

VOL.42, 1982

I ca

on November 10, 2019 by guest

http://jvi.asm.org/

[image:8.505.265.459.251.530.2]

464 GIDONI ET AL.

consist mainly of oligomers which were

ob-served boundtosequencesattheviral origin of replication. To reconcile theirobservationswith

the binding studies reported herein, it can be

suggested that either highly purifiedD-2protein isfunctionally different from 16S T antigen

iso-lated from extracts of SV40-infected cells or

that, although the active binding form of T

antigen is the 5to7S form, the 16S oligomersare

generated through DNA binding. Possibly, T

antigen binds to DNA as monomer or dimer,

whichcould notbeseparated in these sedimen-tation conditions, and then forms tetramers

throughjuxtaposition at its binding sequences.

Although cessation of DNA replication did not

reduce the appearance of 16S T antigen, it is

possible that these oligomers are formed by

binding to cellularornonreplicating viral

chro-matin. Experiments are under way to test

whetherT antigen oligomers can be generated

through bindingtoDNA invitro.

Another area of investigation concerns the nature ofthe 16S oligomers. As the monomer

molecular weight ofTantigen is 82,000(7), the 5

to7Sformprobably consists ofmonomers and possibly dimers, whereas the 16S form, which sediments somewhat more slowly than native

thyroglobulin (molecular weight, 670,000), may

consistofoligomers ranging from tetramers to octamers.Ithas beenreportedthat themajority

of D-2 T antigen present in highly purified, concentrated protein preparations existsas

tet-ramers, as estimated by electron microscopic

analysis (21). Again, it is difficult toassess the

extent towhichpurifiedD-2tetramersresemble

the 16S T antigen observed in crude extracts.

Although it is possible that Tantigen

oligomer-ization in infected cells maybe the result ofits

complexing with other viralorcellularproteins, noevidenceofany specifically

immunoprecipi-tated [35S]methionine-labeled protein associated

with16STantigenwasobserved. It is alsonot inconceivablethatsome16STantigen is, in fact,

a protein-DNA complex. Our experiments do not rule out the existence of short, protected

sequencesof nucleic acid.Experimentsto

deter-minewhether thereareotherproteinsornucleic

acidfragmentsassociatedwith16STantigenare

inprogress.

SV40Tantigenhas beenshowntobindto(6,

9, 14, 18, 20) and stabilize (22)an

immunopre-cipitation-resistant complex with a

53,000-mo-lecular-weighthostprotein foundin transformed

cells. This complex was shown to sediment

morerapidlythan the 16STantigen form(18,20,

28). Althoughit has been suggestedthat 16ST antigenmaybeanintermediatein the formation

of the22S T-53,000-molecular-weight host

pro-teincomplex, thereare substantialquantitiesof

16S T antigen in lytically infected cells which

have considerably smaller amounts of the

53,000-molecular-weight

protein (10, 20).

The

properties of theTantigenthatis

complexed

to

the

53,000-molecular-weight

host

protein

are

dif-ferent fromthose ofthe 16S form because the

formercanbindtoDNA sequencesatthe viral

origin (Scheller et al., submitted for

publica-tion), whereas the latter doesnot.

The existence ofmultipleforms ofT

antigen

differing in their state of

oligomerization

has beenreported by several othergroups.

Although

minordiscrepanciesamongsomeofthe

reported

studies exist, salientfeaturesofthe

phenomena

canbesummarized asfollows.

(i)

In

lytic

infec-tion the Tantigenoligomers sedimentat approx-imately 16S, whereas intransformed cellsthere are still heavier 22S T antigen

complexes

con-taining the

53,000-molecular-weight

nonviral T antigen. (ii) Pulse-labeling studies suggest that themonomericor5to7S form isformed before

the heavier forms and therefore may be their precursor. (iii) tsA mutant T antigens do not

form oligomers at the restrictive temperature.

(iv) Therapidly sedimenting T

antigen

is more

extensively

phosphorylated

than themore slow-ly sedimenting forms. In

experiments

not de-scribed in this paper we have confirmed all of these observations except for the

fourth,

in which wedidnotobservea

significant

difference

in the degree of

phosphorylation

of the two

sedimenting forms ofT antigen. However, our

dataareconsistentwith the

pulse-labeling

stud-ies in thatwehavepreviouslyshown thatnewly synthesizedTantigen bindsmore

efficiently (23,

28) and tightly (23) to DNA than do older T antigen molecules. We are currently

exploring

thepossibilitythatvariousmonoclonal

antibod-ies whichreactwithsubpopulationsofTantigen

thatare active in DNAbinding (Schelleretal.) preferentially immunoprecipitate the 5 to 7S form.

Biological andbiochemical studies of SV40T antigen suggestthat this viral geneproductisa

multifunctional protein. Although the different functions may be due to different domains or

post-transcriptionally modified states (e.g.,

phosphorylation)of the T antigenpolypeptide,it is also possible that T antigen may exhibit het-erogeneity through various degrees of oligomer-ization. Theobservation that the 5 to 7S and 16S forms differ in their DNA-binding properties may be related to different functions of the protein. We can speculate that, although the newlysynthesized 5 to 7S T antigen form is the active viralorigin-binding form, the subsequent-ly formed more rapidsubsequent-ly sedimenting T antigen

oligomers may (i) bind nonspecifically to DNA and play a role in the cellular DNA replication function of Tantigen; (ii) have a function unre-latedto DNAbinding, such as association with

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

DNA BINDING OF SV40 T ANTIGEN OLIGOMERS 465

the53,000-molecular-weight nonviral T antigen,

or adenosine triphosphatase activity; or (iii) be

inactive forms of T antigen representing an

additional mode of autoregulation of the

avail-ability of functionalT antigen.Weare currently

carryingoutexperimentsto testtheseand other

possibilities.

ACKNOWLEDGMENTS

Studies carriedoutin Israelwerefundedbyagrantfrom the

U.S.-Israel BinationalFund to C. P.;those carried outin the

United Statesweresupported byU.S. PublicHealth Service grantCA 26905toC.P.

R. Pollack and Y. Gronerarethanked for critical reviewof

this manuscript.

ADDENDUMINPROOF

While this paper was submitted for publication, Bradleyetal.(M.K. Bradley,J. D.Griffin,andD. M.

Livingston, Cell 28:125-134, 1982) reported that the

slowly sedimentingform of Tantigencanberesolved

into 5.5Sand7Sspecies. Theyhave reportedthat of

thetwoforms onlythe7S speciesis activeinbinding toDNA.

LITERATURECITED

1. Bjursell,G. 1978. Effectsof2'-deoxy-2'-azidocytidineon

polyomavirus DNAreplication: evidence forrolling

cir-cle-typemechanism. J.Virol. 26:136-142.

2. Brugge,J.,andJ.Butel. 1975.Involvementof the simian

virus 40geneA function in the maintenanceof

transfor-mation. J.Virol. 15:619-635.

3. Carroll, R.B.,L.Hager, and R. Dulbecco. 1974. Simian

virus 40TantigenbindstoDNA. Proc. Natl.Acad. Sci.

U.S.A.71:3754-3757.

4. Cowan, K., P. Tegtmeyer, and D. D. Anthony. 1973.

Relationship of replication and transcription of simian

virus 40 DNA. Proc. Natl. Acad. Sci. U.S.A.

70:1927-1930.

5. Fanning, E., B.Nowak, and C.Burger. 1981.Detection

and characterizationofmultipleforms of simianvirus 40

largeTantigen. J. Virol.37:92-102.

6. Greenspan, D. S.,and R. B. Carroll. 1981.Complex of

simian virus 40 large tumour antigen and 48000-dalton

host tumourantigen. Proc.Natl.Acad. Sci. USA

78:105-109.

7. Griffin, J. D., S. Light, and D. M. Livigston. 1978.

Measurements ofthe molecularsize of the simianvirus 40

largeTantigen.J. Virol. 27:218-226.

8. Griff,J. D., G. Spangler,and D. M. LivIngton.1979.

Proteinkinaseactivityassociatedwith simian virus40T

antigen. Proc.Natl. Acad. Sci. U.S.A. 76:2610-2614.

9. Gurney, E. G., R. 0. Harrison, and J. Fenno. 1980.

Monoclonalantibodiesagainstsimianvirus 40 Tantigens: evidencefor distinct subclassesoflargeTantigenandfor

similarities amongnonviralTantigens. J. Virol.

34:752-763.

10. Harlow, E., D. C. Pim, and L. V. Crawford. 1981.

Complex of simian virus 40 large T antigen and host

53,000 molecular-weight proteininmonkeycells. J.Virol.

37:564-573.

11. Kilton,L.J.,M.Bradley,C.Mehta,and D. M. Livingston.

1981.Rapidand sensitivequantitative immunoassay for

thelargesimianvirus40 Tantigen.J. Virol.38:612-620.

12. Khmura,B.,and R.Dulbecco.1973. A temperature-sensi-tive mutant of simian virus 40 affecting transforming

ability.Virology52:529-534.

13. Kuchino,T.,andN.Yamaguchi.1975.Characterizationof

Tantigenin cells infectedwith atemperature-sensitive mutant of simian virus40. J.Virol. 15:1302-1307.

14. Lane,D.P.,andL. V.Crawford.1979. Tantigenis bound

toahostproteinin SV40-transformedcells.Nature

(Lon-don)278:261-263.

15. Livingston,D.M.,I.C.Henderson,andJ.Hudson.1974. SV40Tantigen:partial purificationandproperties.Cold SpringHarborSymp. Quant.Biol.39:283-289.

16. Mardn, R. G., and J.Y. Chou. 1975. Simianvirus 40 functionsrequiredfortheestablishment andmaintenance ofmalignanttransformation.J.Virol.15:599-612.

17. Mardn,R.G.,V.P.Setlow,A. B.

Chepelisky,

R.Seif,A.

M.Lewis, Jr.,C. D.Scher,C.D.Sties,andJ.Avila. 1979. Rolesof the Tantigensintransformationby SV40.Cold SpringHarborSymp.Quant.Biol. 44:311-324. 18. McCormick, F.,and E.Harlow. 1980. Association ofa

murine53,000-daltonphosphoproteinwith simian virus 40

largeTantigenin transformedcells. J. Virol.34:213-224. 19. McKay,R.1981.Bindingofasimian virus 40 Tantigen

relatedproteintoDNA. J.Mol. Biol. 145:471-488. 20. Melero,J.A.,D.T.Stltt,W. F.Mangel,and R.B.Carroll.

1979.Identification ofnewpolypeptide species(48-54K)

immunoprecipitatedbyantiserumtopurifiedlargeT anti-genandpresentin SV40infectedand transformedcells. Virology93:466-480.

21. Myers, R. M., R. C. Williams, and R. Tjian. 1981.

Oligomeric structureofan SV40T antigenin freeform

andasboundto SV40DNA. J. Mol.Biol. 148:347-353. 22. Oren, M., W.Maltzman,and A. J.Levine. 1981.

Post-translationalregulationofthe 54K cellulartumourantigen

in normaland transformedcells. Mol. Cell Biol. 1:101-110.

23. Oren, M.,E.Winocour, andC.Prives. 1980.Differential affinities ofSV40T antigenfor DNA. Proc. Natl. Acad. Sci. U.S.A. 77:220-224.

24. Osborn,M.,and K.Weber.1975. SV40:Tantigen,theA

function andtransformation. ColdSpringHarborSymp.

Quant.Biol. 39:267-276.

25. Osborn,M.,andK. Weber.1975.Simianvirus 40geneA function and maintenance of transformation. J. Virol. 15:636-644.

26. Ozer, H. L., and P. Tegtmeyer. 1972. Synthesis and

assemblyofSV40 H.Synthesisofthemajorcapsid protein

and itsincorporationintoviralparticles.J.Virol.9:52-60. 27. Potter,C.W.,B. C.McLaughlin,andJ.S. Oxford.1969. Simian virus40-inducedTandtumorantigens. J. Virol. 14:574-579.

28. Prives,C.,Y.Beck,D.Gidoni,M.Oren,and H.Shure.

1979. DNAbindingandsedimentationpropertiesofSV40

tumour antigenssynthesized invivo and in vitro. Cold

SpringHarborSymp.Quant.Biol. 44:123-130. 29. Prives, C., Y. Beck, and H. Shure. 1980. DNA binding

properties of simianvirus 40 T-antigens synthesized in vivoand in vitro. J. Virol.33:689-696.

30. Prives,C.L.,and H.Shure.1979.Cell-freetranslationof simian virus 40 16S and 19S L-strand-specific mRNA

classes tomajor VP-1 and minor VP2 and VP3 capsid proteins.J.Virol.29:1204-1212.

31. Reed,S.I.,G. R.Stark,andJ.C.Alwine.1976. Autoregu-lationofsimian virus40 gene AbyTantigen.Proc.Natl.

Acad.Sci. U.S.A. 73:3038-3087.

32. Rigby,P. W.J.,M.Deckman,C. R.Rhodes,and P.Berg.

1977. Labelling deoxyribonucleic acid to high specific

activityin vitrobynicktranslationwith DNApolymerase.

J. Mol. Biol. 113:237-251.

33.

Rundeil,

K.,J.K.Collins,P.Tegtmeyer,H. L.Ozer,C-i. Lai, and D. Nathans. Identificationof simian virus 40 proteinA.J.Virol. 21:636-646.

34. Shalloway, D., T.Kleinberger, D. M. Livingston. 1980. MappingofSV40DNAreplicationorigin region binding

sites for theSV40Tantigen byprotectionagainst

exonu-clease IIIdigestion.Cell20:411-422.

35. Smith, A.E.,R.Smith,andE.Paucha. Extractionand

fingerprint analysisofsimian virus 40largeandsmall

T-antigens.J.Virol.28:140-153.

36. Tegtmeyer,P.1972.Simianvirus40deoxyribonucleicacid synthesis:the viralreplicon.J.Virol.10:591-598. VOL.42, 1982

on November 10, 2019 by guest

http://jvi.asm.org/

466 GIDONI ET AL.

37. Tegtmeyer,P.1975.Function ofsimian virus 40geneAin

transforming infection. J. Virol. 15:613-618.

38. Tegtmeyer,P., K.Rundell,J.K.Collins.1974. Modifica-tion of simian virus 40 protein A. J. Virol. 21;647-657. 39. Tegtmeyer,P., M. Schwartz, J. K. Collins, and K. Rundell.

1975. Regulation oftumourantigensynthesis by simian virus40geneA. J.Virol. 16:168-178.

40. Tjhan, R. 1978. The binding siteonSV40 DNA foraT

antigen-related protein. Cell 13:165-179.

41. T3ian, R., and R. Robbins. 1979. Enzymatic activities

associated with a purified simian virus 40 T antigen-related protein. Proc. Natl. Acad. Sci. U.S.A. 76:610-614.

42. Tooze, J. (ed.). 1980. DNA tumor viruses: molecular biology oftumourviruses, 2nded.,part2. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y. 43. Voller, A., A.Bartlett,D. E. Bidwell,M. F.Clark,and

A. N. Adams.1976.The detection of viruses by

enzyme-linked immunosorbent assay (ELISA). J. Gen. Virol.

33:165-167.

J.VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/