Polyomavirus and simian virus 40 large T antigens bind to common DNA sequences.

Copyright ©1984, American Society for Microbiology

Polyomavirus and Simian Virus 40 Large

T

Antigens

Bind

to

Common DNA

Sequences

BETSY J. POMERANTZAND JOHNA. HASSELL*

Department ofMicrobiology andImmunology, McGill University, Montreal, Quebec, Canada H3A 2B4

Received 2 October1983/Accepted 5December1983

The large T antigens of polyomavirus and simian virus 40 (SV40) recognize and bind to specific, noncoding DNAsequenceswhicharelocated between thebeginning of the early and late transcription units intheir respectivegenomes.Eachlarge T antigen bindstomultiple sites within this intergenicDNAstretch. Polyomavirus large T antigen bindsto atleasttwosites within its DNA,andSV40largeTantigen bindsto three sites within SV40 DNA. Comparison of the DNA sequences which comprise the binding sites in polyomavirus DNA orthose which make up the binding sites in SV40 DNA has led to recognition ofa

common sequence, -GAGGC-, which is repeated within each large-T-antigen-binding site. We tested the hypothesis thatrepeatsof this pentanucleotide form therecognition-bindingsiteforpolyomavirusandSV40 large T antigen. This was accomplished by measuring the binding of each large T antigen to both polyomavirus andSV40DNAandto synthetic DNA substrateswhich did ordidnotcontainrepeats ofthe -GAGGC- sequence. Polyomavirus large Tantigen bound to specific fragments ofSV40 DNA, and SV40 large T antigen bound with specificity topolyomavirus DNA. In each case, the DNAfragments bound by the heterologous large T antigen were the same as those bound by the homologous large T antigen. Moreover, polyomavirus and SV40 large T antigen only bound to synthetic DNA substrates which contained repeats of the pentameric sequence. This synthetic DNA also competed effectively withnative polyomavirusorSV40 DNAas asubstrateinbinding reactions withone orthe otherlargeTantigen. These results ledustoconclude that repeatsof the-GAGGC- sequenceform therecognition-binding site for both polyomavirus and SV40large T antigen.

Polyomavirusandsimian virus40(SV40),twomembersof

the papovaviruses, share a common virion architecture,

genome structure and organization, and replication cycle.

With some exceptions, this synonomy extends tothe

num-ber,molecularweights, amino acidsequences, and functions of the viral-encoded proteins (13, 20, 42). Nonetheless, important differences distinguish these viruses. They

repli-cate in radically different hosts; polyomavirus multiplies in murine cells, and SV40 multiplies in simian cells. The

products of their transcription units are not all identical;

polyomavirus encodes one moreearlygeneproduct(middle Tantigen) than does SV40. There are sequencedifferences within the intergenic, regulatory regions of their genomes

(20).The latterincludenot only alterationsinprimaryDNA structure,but alsodifferencesin theorganizationof the viral controlelementssuch astheorigin forDNAreplication (31),

the early and late promoters (15, 24), and the number and

locationofbinding sites for regulatory proteins (15, 34). We havebegun todefinethe borders and toidentify the

impor-tant DNA sequences within these intergenic regulatory

regions in polyomavirus andhavecompared their sequence andorganizationtosimilarcontrolelements inother papova-viruses to derive general models for the initiation and

regulation of DNAreplicationandtranscription (24, 31, 34). Because both viruses make extensive use ofhost cell

ma-chinery to implement these processes, we anticipate that what we learn about replication and transcription from studies of papovaviruses can be generalized to mammalian cells.

Thereplicationofcyclesofpolyomavirus and SV40 occur in twotemporal stages. Thefirst, defined as the early phase,

begins soonafterinfection and extendsto the onsetof viral DNA replication. During this time the viral early

transcrip-*Corresponding author.

tion units are expressed to yield theearly proteins. Polyoma-virus encodes three early proteins, large, middle, and small Tantigen, whereasSV40 codes for two early proteins, large

and small T antigen (20). Thesynthesis of viral DNAsignals the beginning of the late phase of infection. During this phase, the late transcription units areexpressed to yield the three structural proteins of polyomavirus and SV40 VP-1,

VP-2, and VP-3 (20). Now the relative synthesis of tran-scripts from the early transcription unit is drastically re-duced. Both the initiation of viral DNA replication and the repression of early mRNAsynthesis at late times is mediated by large T antigen (7, 8, 44, 48). The mechanism by which large T antigen performs these functions has not been elucidated in equal detail for both viruses. However, it is verylikelythatin both caseslargeTantigenmustphysically

interact with specific viral sequences to effect both replica-tion and early transcription (41).

The large Tantigens ofpolyomavirus andSV40are both nuclear phosphoproteins with calculated molecularweights of87,991 (785 amino acids) and 81,632 (708 amino acids),

respectively, althoughdirect measurements of their molecu-lar weights have led to higher estimates ofbetween 88,000 and 100,000 (20). Both proteinsexist as multiple molecular species (2, 11, 12, 19, 22, 27, 36). A fraction of the oligomeric

species ofSV40 large T antigen, and perhaps of

polyoma-virus large T antigen, contain a cellular-encoded

phospho-protein, P53, also referred to as nonviral tumorantigen or Tau antigen (11, 19, 22, 26). The biological roles of these variousforms oflarge T antigen are not known with certain-ty, but they differ in their extent of phosphorylation and possess different activities in vitro (2, 11, 17, 36). One

activityof the largeTantigenencodedbypolyomavirusand

SV40 is the capacity to hydrolyze ATPto ADP and

Pi

(14, 16, 21, 49). Anotheractivity is the capacity to bind to DNA (4, 14). Polyomavirus large T antigen binds to at least two 925

on November 10, 2019 by guest

http://jvi.asm.org/

926 POMERANTZ AND HASSELL

sites within polyomavirus DNA (34) (Fig. 1). One of these

sites overlaps with DNA sequences which form part of the

replication origin, whereas the other straddles sequences

importantfor earlymRNA synthesis (15, 24, 31). SV40large T antigen binds to three, closely spaced sites within SV40

DNA which occupy a region 120 to 180 base pairs (bp) in length (9, 40, 43, 45-47) (Fig. 1). These large-T-antigen-binding sites overlap sequences which comprise the early promoter and origin for DNA replication (20) (Fig. 1). Experiments with SV40 DNA and a large-T-antigen-related

protein, the D2 protein, have shown that the guanine resi-dues in the binding sites which are part of the sequence

-GAGGC-areprotected by the D2 protein from methylation withdimethyl sulfate (47). This pentanucleotide is repeated in two of the SV40 large-T-antigen-binding sites (Fig. 1).

Repetitionof this experimentunder avariety of experimen-talconditionswithauthenticSV40 large Tantigen has led to

nearlyidentical conclusions (9).

Inspectionofthe DNAsequences comprising the

large-T-antigen-binding sites in

polyomavirus

DNA reveals that the

related sequence -A/TGAGGC- is repeated in both of its binding sites (Fig. 1). The conservation of

homologous

hexameric and pentameric repeats within the large-T-anti-gen-binding sites in polyomavirus andSV40 DNA suggests

that these sequences are functionally important and might

serve as the recognition-binding sites for both large T antigens (34). To test this hypothesis, we measured the

bindingofpolyomavirusorSV40 largeTantigen to the DNA of the other andtosyntheticDNApolymerswhich contained repetitions of the -GAGGC- sequence. The results ofthese experiments substantiate the contention that repeats ofthe

pentanucleotideform the coreofthepolyomavirus and SV40

large-T-antigen-binding sites.

MATERIALS AND METHODS

Cells and viruses. Cells were grown onplastic disheswith Dulbecco modified Eagle medium supplemented with 10%

(vol/vol) fetal bovine serum and antibiotics and maintained

Polyomavirus

TATA Early mRNAs

*

AT 5292/1

I I I I _

5130 5180 5230 5280

T antigen binding sites

--D --Om --O

II

I I I -

-I

___- -

-Functional replitation origin

I

p

I

Core

SV40

I 2

250 200

TATA Early mRNAs

I I I5

rw

150 100 S0 1/5243 5193 51

Tantigen binding sites

~ -00

Functional replication origin Auxiliary I Core

FIG. 1. Polyomavirus andSV40large-T-antigen-bindingsitesrelative to other controlsequences.Theregulatory regionsofpolyomavirus andSV40areshown. The nucleotidenumberingscheme of Soedaetal.(42)wasused forpolyomavirusandthat of Buchman et al.(3)wasused

forSV40. The solid diamond represents thejunctionbetween thefirst and last nucleotide in the DNA sequence.The -TATA- consensus sequenceand the5'termini of the abundantearlymRNAsareshownbythe solid boxes(15, 20).Thedirection ofearlytranscriptionisfrom

lefttorightasshownby thearrow onthe solid boxrepresentingtheearlymRNAs. Theputativeinitiationcodons for translation of theearly mRNAsareboxed. Thelarge-T-antigen-binding sitesareillustrated byopenboxes. The boundaries of thepolyomavirus large-T-antigen-bindingsitesweretaken from Pomerantzetal.(34)and those ofSV40weretaken from DeLucia etal.(9).Thearrowsabove and belowthe

large-T-antigen-bindingsites refertothe -GAGGC-sequence,and theirpositionaboveorbelowtheopenbox refer to their locationononeor theother DNA strand. The minimal DNAsequenceswhich functionasreplication originsarealso shownasopen boxes. For eachvirusthere

are a minimumoftwo sequence domains comprising the functional replication origin. For polyomavirus these are referred to as ,3 and

core,whereasforSV40 theyarereferredtoasauxiliarysequencesandcore.The borders of these sequencesaretaken from Muller et al.(31) forpolyomavirus and fromMyersandTjian(33), DiMaio and Nathans(10), andBergsmaetal.(1)forSV40.

I I I I I

50 100 150 200 240

300 43

---s

J.VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:2.612.131.489.281.624.2]

at 37°C in a humidified CO2 atmosphere. Polyomavirus stocks were prepared by infection of primary baby mouse

kidney cells at low multiplicity (0.01 to0.1 PFU per

cell).

Infected cells were harvested 10 to 14 days postinfection. Stocks ofSV40were

prepared similarly by

infection ofCV-1 cells. Viruswas releasedfrom infected cellsbythree

cycles

offreezingand thawing.

Construction of recombinant plasmids. pPSVE1 is

com-prised of pBR322 sequences, the SV40HindlIl C fragment

(SV40 nucleotide 1047 to5171), and the

polyomavirus HphI

(nucleotide 154) to EcoRI (nucleotide 1560) fragment (see Fig. 2). The numberingsystemsusedarethoseproposed by

Soedaet al. (42)and Buchmanet al. (3)for the A2 strain of

polyomavirus and the 776 strain ofSV40, respectively. We

cite nucleotides from the clockwise direction on the

poly-omavirus and SV40

physical

maps. The SV40sequences in

pPSVE1werejoinedtothe

polyomavirus

sequences insuch

awayas tojuxtapose the SV40earlypromoterimmediately adjacenttothe polyomavirusDNA which encodes the

early

proteins. The details concerning the construction of this

recombinant plasmid will be described elsewhere.

Impor-tantly, pPSVE1 contains the natural sites of

binding

for

SV40largeTantigen

(within

thebordersof theSV40 HindlIl

C fragment)butdoes notcontain the natural sites of

binding

for the polyomavirus large T

antigen.

The methods usedto construct pPH1-8 (25) and

pSVOl (33)

have been described (see Fig. 2).

Isolation of DNA and its modification. Recombinant

plas-midDNAswere isolated fromEscherichia coli strainHB101

orDH1 and

purified by

CsCl

density centrifugation.

Restric-tion enzyme

digestions

were

performed

in accordance with

the

specifications

of the manufacturer. To end label the

DNAfragments, 100 ngof restriction

enzyme-cleaved

DNA was incubated in a volume of 50 ,ul with 10 mM

Tris-hydrochloride (pH 7.6), 5 mM MgCl2, 1 mM

dithiothreitol,

25 p.Ci of the appropriate a-32P-labeled

deoxynucleotide

triphosphate (3,000

Ci/mmol),

and 5 U of the Klenow

fragment ofE. coli DNA polymerase I for 1 h at

15°C.

The

reaction was terminated by phenol

extraction,

and the aqueous phase was

chromatographed

on a

Sephadex

G-50

column to separate the

unincorporated

deoxynucleotide

triphosphates from the labeled DNA

fragments. Typically,

the specific activity of the labeled DNA

fragments averaged

108 cpm/,ug.

Synthetic

oligodeoxynucleotides

were

phosphorylated

and ligated foruse asDNAsubstrates in the

binding

assays. From 1 to 4

jig

of DNA was labeled with 50

p.Ci

of

[y-32P]ATP

and 5 U ofT4kinase in abuffer

containing

70 mM

Tris-hydrochloride (pH 7.6), 10 mM MgCl,, and 5 mM

dithiothreitol. After 1/2 h at 37°C, unlabeled ATP

(final

concentration, 0.5 mM) and 5 U ofT4 kinase were

added,

and thereactionwascontinuedfor another 1/2 hat

37°C.

The

phosphorylatedlinkerswere

ligated

for12to16 h at 15°Cin 50 mM

Tris-hydrochloride (pH

8.0)-10 mM

MgCl2-13

mM

dithiothreitol-1mMATPwith 10 U ofT4

ligase

in avolume

of100

[L.

Preparation ofnuclear extractsfrominfected cells. Nuclear

extracts were prepared from 3T6 cells infected with

poly-omavirus or CV-1 cells infected with SV40

essentially

as describedby McKay (28). From 0.5 x

109

to 1.0 x

109

cells

growingon the surface of plastic petri dishes were infected with virus at a multiplicity of infection of about 10. The cells

wereharvested by scraping (40 to 46 h postinfection at

37°C)

and then washed twice with ice-cold phosphate-buffered salinesolution. The cellswereresuspended in

lysis

buffer (10

mM Tris-hydrochloride, pH 7.4, 10 mM NaCl, 1.5 mM

MgC92,

0.5 mM dithiothreitol, 100

p.g

of phenylmethylsul-fonyl fluoride per ml) ata density of5 x

107

cells per ml.

Theywere allowedtoswell at4°Cfor10 min and thenlysed

by 10to 15 strokes with atype B pestle in a glassDounce

homogenizer. The nuclei were pelleted by centrifugation

(2,000 rpmfor10 min at4°C)andwashed inlysisbuffer. The

nucleiwerethenresuspended in nuclearextraction buffer(10 mM Tris-hydrochloride, pH 8.0, 150 mM NaCl [for

SV40-infected CV-1 cells] or 300 mM NaCl [for polyomavirus-infected 3T6 cells] plus 0.5 mM dithiothreitol, 100 ,ug of phenylmethylsulfonyl fluoride per ml) and incubated at4°C

for5 to 15 min. The nucleiwere pelleted bycentrifugation,

and the supernatantwasusedas a sourceoflargeTantigen. These preparations were storedat -70°C until use.

Immunoassay. The conditions of the

immunoassay

were

essentially those described

by

McKay

(28).

Briefly,

reac-tions were carried out in a 1-ml volume

containing

20 mM

NaPO4

(pH 7.0),2 mM

dithiothreitol,

0.01%

(wt/vol)

bovine

serum

albumin,

0.1 mM EDTA, 0.05%

(vol/vol)

Nonidet

P-40,40 ,ugof

phenylmethylsulfonyl

fluoride per

ml,

and3.0%

(vol/vol) dimethyl sulfoxide.From 1to10 ngof labeled DNA

fragmentsandvariousamountsofthe

large-T-antigen

prepa-rationswerealso included. Aftera 1-hincubationat

25°C,

2 ,ugofunlabeledratcellular DNAwas

added,

andincubation

continuedfor 20 min at

25°C.

The inclusion of

nonradioac-tive DNA reduces the

nonspecific binding

of

large

T

antigen

to the labeled DNA. The

large

T

antigen-DNA complexes

wereimmunoprecipitated by

sequential

incubation with anti-bodies directed

against

the viral

large

T

antigens

and then with Formalin-fixed

Staphylococcus

aiureius

cells. These

reactions were

performed

at

25°C

for 20 min in each

in-stance.

Polyclonal

antisera directed

against polyomavirus

large T antigen were obtained from ascites fluid of brown

Norwegian rats

bearing

tumors induced

by

polyomavirus-transformed rat cells. Hamster antiserum directed

against

SV40largeTantigenwas obtainedfrom the National

Insti-tutesofHealth(serum-ID = 81 x0000001). Neither

antiser-um was purified before use. The

immunocomplexes

were then washed twice with 1 ml of 10 mM

Tris-hydrochloride

(pH 8.0-150 mM NaCI-0.5% Nonidet P-40.

Finally,

the DNA was released from the

immunocomplexes

with 1% (wt/vol) sodium dodecyl sulfate-10 mM EDTA and depro-teinized with

phenol

and

chloroform/isoamyl

alcohol

(24:1

[vol/vol]), and a portion was electrophoresed through an

agarose or

polyacrylamide gel;

1.5%

(wt/vol)

agarose

gels

and 12%

(wt/vol) polyacrylamide gels

were used to

analyze

the DNA.

Electrophoresis

was

performed

in 89 mM

Tris-hydrochloride

(pH 8.0)-89 mM boric acid-2 mM EDTA.

Before

autoradiography,

agarose

gels

were immersed in ethanol for 30 min and dried under vacuum with

heating.

Polyacrylamide

gelswere notdried before

autoradiography.

Generally,

gels wereexposed to Kodak XAR-5film for6to 36 h in casettes with Dupont

Lightning-Plus

intensifying

screens.

RESULTS

Weemployedtheimmunoassay described by McKay (28)

to measurethe bindingoflargeTantigen presentin nuclear

extracts from

lytically

infected cells to end-labeled

frag-mentsofpolyomavirus orSV40 DNA. This assay makesuse

of the fact that DNA bound to a

particular

protein

can be

separated from free DNA by immunoprecipitation with

antibody

specific

to that protein. In

brief,

defined radiola-beled DNA

fragments

were reacted with nuclear prepara-tions oflargeT antigen,and thefragments bound

by large

T

antigen

were

immunoprecipitated

withanantiserumdirected

on November 10, 2019 by guest

http://jvi.asm.org/

928 POMERANTZ AND HASSELL

V

O3

75

e4~

VP=;:

;

~~~~~53

as EcoRN 10

9so

/2

oso0)

EcoR

11

\6

/(5092) 3 \

225 7

a8

PsVO1

6 249

48713bp_

\~~~~~~~~~~~~~~~

11

X /270

1743

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

against large T antigen and Formalin-fixed S. aureus cells. The immunocomplexes were washed, and the eluted DNA

was analyzed by gel electrophoresis and autoradiography. Only DNA fragments which were boundspecifically by the large T antigen were displayed on the autoradiogram. For convenience, and to simplify the interpretation of the

re-sults, we used recombinant plasmid DNAs which were

known to harbor the large-T-antigen-binding sites as

sub-strates inthe binding reactions. Their structures are shown

in Fig. 2togetherwith appropriate restriction endonuclease cleavage maps. In addition, the location of the

large-T-antigen-binding sites, themajorcap site, and the direction of transcription of the early mRNAs are shown. pPH1-8

con-tains the two known polyomavirus large-T-antigen-binding

sites (34) within the 604-base pair (bp) Hinfl 4 fragment, which is composed of only polyomavirus DNA (Fig. 2).

pPSVOlandpSVE1 containthe threeSV40 large-T-antigen-binding sites (33) within the borders of the Avall-HindlIl 2 and 4fragments, respectively (Fig. 2).

Specific bindingofpolyomavirusandSV40largeTantigen

to each other's DNAs. To test thehypothesis thatrepeats of the sequence -GAGGC- arerequired for the specific binding of polyomavirus and SV40 large T antigen to DNA, we

measuredthe binding of each large T antigen to its own DNA and to theheterologous DNA. We employed this approach because each of the large-T-antigen-binding sites in the genomes of these viruses contains repeats ofthe aforemen-tioned sequence(Fig. 1). Wereasoned that ifrepeats of this sequence are important forbinding oflarge T antigen then eachspecies of largeTantigenought to be capable of binding toboth DNA substrates. Initially, wemeasured the binding

ofpolyomavirus largeTantigen present in a nuclear extract from infected 3T6 cells to end-labeled DNA fragments

obtained by Hinfl hydrolysis ofpPH1-8 DNA (a recombi-nantplasmidDNAwhich containspolyomavirus sequences) (Fig. 2) and to the end-labeled fragments obtained after cleavage of pSVOl DNA (a recombinant plasmid DNA whichcontains theSV40large-T-antigen-binding sites) (Fig. 2) withAvall and HindIII (Fig. 3A and B). We previously

showed thatpolyomavirus largeTantigen binds specifically

totheHinfl4fragment of pPH1-8DNA(34), a resultwhich is confirmed here (Fig. 3A). When polyomavirus large T

antigen was incubated with the end-labeled AvaIl-HindIll fragments of pSVO1 DNA, the second largest fragment in themixture was selectively bound(Fig. 3B). Thisfragment contains the three SV40large-T-antigen-binding sites (Fig. 2).

One of the problems encountered with this assay is the

immunoprecipitation of DNA fragments other than those whichcontainthelarge-T-antigen-binding sites(Fig. 3A and

A B

DNA Hinf L,'pPH1- 8 Ava I1+Hina 111ir)SVO|

(ng Iml) 1 1 5 5 1 1 5

T-antigen Polyomavirus PolyornaviIr.s l

(ul1) 25 50 25 50 25 50 f5 50

1W

-1 - 1 _ _

2 0 -* 5

3 40

8

9 ~~~~~~~4

10.11 5

12,13 A* 14.15S

16-20 8-11 i

M 1 2 3 4 M 1 2 3 4

FIG. 3. Immunoprecipitation ofpolyomavirus and SV40 DNA fragments after reaction with nuclear extract from polyomavirus-infected cells. (A) pPH1-8 DNA was cleaved withHinfl (see Fig. 2A), and thefragments were endlabeled and usedas substrate in binding reactionsattheconcentrations shown in thefigure. Several concentrations of nuclearextractfrom polyomavirus-infected 3T6 cellswereassayed. The marker lane (M)containsaportion ofthe original DNA substrate (pPH1-8 cleaved with Hinfl). Fragment 4 contains the previouslymappedpolyomavirus large-T-antigen-bind-ing sites(seeFig.1and 2).The driedagarosegelwasexposedtofilm for18 h. (B) Nuclearextractfrompolyomavirus-infected 3T6cells was reacted with end-labeled fragments generated by AvaIl and

HindIll cleavage of pSVO1 DNA (see Fig. 2). M represents a portion of the

AvaIIIHindIII

digestion products of pSVO1 DNA usedassubstrate. Fragment 2contains the previouslymapped SV40 large-T-antigen-binding sites (Fig. 1 and 2). The driedagarosegel wasexposed tofilm for18 h.

B). These fragments are immunoprecipitated to a lesser extent than thosewhich containthebinding sites forlarge T

antigen (Fig. 3A and B). The largest fragments are usually

preferentially precipitatedunder these conditions (see

espe-cially Fig. 3A lanes 3 and 4and Fig. 3B lanes3 and 4). We suspect that thisoccurseither because large T antigen binds

nonspecifically to DNA fragments or because the protein

bindsspecificallytoweakrecognition-binding sitesin DNA. In addition, we have found that many preparations of

antiserum contain DNA-binding proteins. Their interaction

with the labeled DNA fragments also contributes to the

background bindingweobserved.

To determine whether SV40 large T antigen would bind

FIG. 2. Structures and restriction endonucleasecleavage mapsof recombinantplasmidDNAs. TherecombinantplasmidpPH1-8contains polyomavirussequencesbetweentheHindIll cleavagesitesatnucleotides3918 and 1656(clockwiseon thepolyomavirus DNAmap)inserted into theuniqueHindIll site of pBR322.The large openbox representspolyomavirusDNA; thethin line represents pBR322 sequences. Small arrowsdesignate sites ofHinflcleavage, and the two larger arrows denote those Hinfl sites which flank the previously mapped large-T-antigen-binding region. Sizes of restriction fragments (inbp) are notedaround the periphery of the plasmid. Therelativesize of each fragment isindicated bythe number within the circle. The symbols above the plasmid represent features of polyomavirus DNA which reside between the early andlate transcription units. The open triangle denotes the area in which DNA replication initiates, the closed box represents the sequencesboundby the polyomaviruslarge Tantigen,and the open arrow depicts the site ofinitiationandthedirectionof transcription of the early viral mRNAs. pSVOlcontainsthe EcoRIIG fragmentofSV40 DNA(nucleotide5092 to 160)inserted into the uniqueEcoRI site of pBR322DNA(nucleotide4362/1). The SV40large-T-antigen-bindingsites arecontained within the EcoRII G fragment. The hatched area represents SV40 sequences. The other symbols are the same as those used above. pPSVE1 contains theHindIlI C fragment ofSV40 DNA (nucleotide 1047 to 5171; thehatched box) and theHphItoEcoRIfragment of polyomavirus DNA (nucleotide 154 to 1560; the open box) cloned between theHindIll (nucleotide29)andEcoRI(nucleotide4362/1)sitesof pRB322 DNA. AHindIll recognitionsite is present at the junction oftheSV40(nucleotide5171) andpolyomavirus (nucleotide 154) DNAsequences. The hatched box represents SV40 DNA, the open

boxrepresentspolyomavirusDNA, and thethinline represents pBR322 DNA. The othersymbols are the same as those used above.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.612.332.547.75.265.2]

930 POMERANTZ AND HASSELL

with specificity to DNA substrates which contain the binding sites for polyomavirus large T antigen, we measured its

capacity tobindto defined fragments of pPSVE1 and pPH1-8 DNA (Fig. 2 and 4A and B). As expected, the nuclear extractcontaining SV40 large T antigen bound with specific-ity to the AvaII-HindIII 4 fragment of pPSVE1 DNA (Fig. 4A).The latter fragment carries the sites of binding forSV40 large Tantigen. In addition, the Hinfl 4 fragment of pPH1-8 DNA, which contains the sites of binding for polyomavirus large T antigen, was specifically bound by SV40 large T antigen (Fig. 4B). These results show that SV40 large T antigen binds with specificity to a DNA fragment which contains the sites of binding for polyomavirus large T antigen.

To better gauge the affinities of polyomavirus and SV40 large T antigens for each other's natural binding sites, we compared the binding of the two large T antigens to the other's DNA under conditions where the concentration of

nuclearextract was limiting and all the other components in the bindrigassay were in excess. To make the comparison

asmeaningfulaspossible, we simultaneously prepared fresh nuclear extracts from 3T6 cells infected with p;lyomavirus and CV-1 cells infected with SV40. In each case, the cells were infected at a multiplicity of 10, and extracts were prepared 48 h postinfection from the same number of cells and were handled identically before they were assayed. A

A B

DNA Ava IX+Hind IIf/pPSVE1 HinfIf/pPH1-8

(ng/mI) 1 1 5 5 1 1 5 5

T-antigen SV40 SV40

(ul) 50100 50 100 50 100 50 100

2,3 ES

44 4

--_*

~~5-7

6

7,889

₉₆

_i10,19!

10

11,12 12.13

13

14,15

14 16

15--17 * 17-20

M 1 2 3 4 M 1 2 3 4

FIG. 4. Immunoprecipitation of polyomavirus and SV40 DNA fragments after reaction with nuclear extract from SV40-infected CV-1 cells.(A)pPSVE1DNAwasdigestedwithAvaIl andHindlll (Fig. 2), and the fragments were end labeled and reacted with nuclearextractfromSV40-infectedCV-1 cells. Theconcentration of DNAand extract used in each reaction is shown in thefigure. M represents asampleoftheAvaIl andHindIll fragmentsofpPSVE1 DNA used as substrate. Fragment 4 contains the SV40

large-T-antigen-binding sites. The dried agarosegel was autoradiographed for 24 h. (B) pPH1-8 DNA was cleaved with Hinfl, and the fragments were end labeled and used as substrate in binding

reactions with nuclear extractfromSV40-infectedCV-1 cells. The concentrations of DNA and extract are shown in the figure. M shows the Hinfl digestion products of pPH1-8 DNA used as

substrate. The fourth largest fragment contains the polyomavirus large-T-antigen-binding sites.Thedriedagarose gelwas autoradio-graphed for24 h.

(.ti) 25 5 f 0

24 3 _ 4 5-7*

8

12,13

-14,15 *

16 20 9

L

M 1 2 3 4

FIG. 5. Immunoprecipitation of polyomavirus DNA fragments after reaction with nuclear extract from polyomavirus- or SV40-infected cells. pPH1-8 DNA was cleaved with Hinfl, and the fragmentswereend labeled and reacted with nuclear extract from either polyomavirus-infected3T6 cells orSV40-infectedCV-1 cells. AfixedconcentrationofDNA but variousamounts of extract were assayed. M represents a sampleoftheHinfI-cleaved pPH1-8DNA used as substrate. The dried agarose gel was exposed tofilm for 18h.

comparison of the capacities of polyomavirus and SV40 large Tantigenstobindto aDNAsubstrate which contained the polyomavirus large-T-antigen-binding sites is shown in

Fig.5. The results showed that bothlargeTantigensbound

totheHinfl 4 fragment of pPH1-8toapproximatelythe same extent.

Wealso compared thebinding of polyomavirus and SV40 large T antigens to DNA fragments which contained the SV40 large-T-antigen-binding sites (viz., pPSVE1 DNA) (Fig. 6). These results showed that both large T antigens bound specifically to the Avall-HindIl 4

fragment

of

pPSVE1 DNA. However,unlike the results obtained previ-ously, the affinity of polyomavirus large T antigen for the DNAfragmentcontainingthe SV40large-T-antigen-binding

sites appeared to be less than that ofSV40largeT antigen

(Fig. 6).

Insummary,the resultsweobtainedstronglysuggestthat

thelarge Tantigens ofpolyomavirus and SV40are capable

of

binding

to each other's genomes with specificity.

More-over, our data are consistent with the hypothesis that the two largeT antigens recognize andbind to the same DNA

sequences.

PolyomavirusandSV40 largeT antigenspecific bindingto synthetic DNA substrates which contain repeats of the se-quence -GAGGC-. To

directly

test the

hypothesis

that re-peatsof the sequence -GAGGC-arerequiredfor thebinding

of polyomavirus and SV40 large T

antigens

to

DNA,

we searched for this sequence among

commercially

available

syntheticDNAlinkers andadaptors. Wenotedthatthe

self-ligation of XhoI linkers

(-CCTCGAGG-)

resulted in the

creation of the -GAGGC- sequence and that trimers of XhoI linkers caused the duplication of this sequence within the molecule separated by3 bp.This arrangement of the penta-nucleotides mimics their topography in the large-T-anti-J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.612.369.508.74.292.2] [image:6.612.75.291.357.575.2]

gen-binding sites of polyomavirus and SV40 DNA. There-fore, we tested the capacity of self-ligated XhoI linkers to serve as a substrate for the binding ofpolyomavirus and SV40 large T antigen. To ensure that the conditions we employedfortheseassayspermitted only specific binding to bescored, we also measured thebindingoflarge T antigen to three other synthetic DNA substrates. These included

self-ligated BamHI (-CGGGATCCCG-), BclI (-CTGATCAG-), andClaI (-CATCGATG-)linkers. Polymersof these linkers

do notform the -GAGGC- repeat. The results of measuring

the binding of polyomavirus large Tantigen to these DNA substrates are shown in Fig. 7. Only the ligated XhoI linkers were specificallybound after reaction with nuclear

prepara-tionsof polyomavirus large T antigen. Moreover, the

larger-molecular-weight polymers were preferentially bound by

polyomaviruslarge T antigen (Fig. 7). The smallest multimer

of the ligated XhoI linkers to be bound was a pentamer. Pentamers contain four repeats of the -GAGGC- sequence per DNA strand, with each repeat separated from the

adjacentrepeat by 3 bp. Inspection of Fig. 7 also reveals that there was virtually no binding of polyomavirus large T

antigen to the ligated BamHI, BclI, or ClaI linkers. This suggests that the background binding which we observed

previously (see Fig. 3A and B) was caused bythe specific binding of polyomavirus large T antigen to weak binding sites in theDNA.

Werepeated this experiment butemployed SV40 large T

antigenin thebindingassay(Fig. 8). Again,the

polymerized

XhoI linkerswereefficiently bound by SV40largeTantigen, but under identical conditions, polymers ofBamHI, BclI,

andClaI linkerswere not(Fig.8).AlsoshowninFig.8Aisa

comparison between the capacities of polyomavirus and SV40large Tantigenstobindtotheself-ligatedXhoIlinkers.

DNA Ava I1+Hind lII/pPSVE1

(naQ/nil) 5 5 5 5

T-antigen SV40 Py

(yI) 25 50 25 50

2,3 *

4 -4 Q

5

7,8-

9- I10-1

1,12k-13 " 14

15-17

M 1 2 3 4

FIG. 6. Immunoprecipitation of SV40DNAfragments after reac-tion withnuclear extractfrompolyomavirus-orSV40-infected cells. pPSVE1 DNA was cleaved with AvaII and HindIII, and the fragments wereend labeledand reacted withnuclearextractfrom eitherpolyomavirus-infected3T6cellsorSV40-infected CV-1 cells. Afixed concentration ofDNA wasreactedwithincreasingamounts of nuclear extract. Mrepresents asample ofthe DNA fragments used as substrate. The dried agarose gelwas exposed to film for 24 h.

FIG. 7. Immunoprecipitation of ligated synthetic linkers after reaction withnuclearextractfrompolyomavirus-infected cells. The linkers were phosphorylated andligatedand thenusedassubstrates in binding reactions with nuclearextractfrom polyomavirus-infect-ed 3T6 cells (30 ,ul perreaction).Theconcentration of linkerDNA was 100ng/ml (lanes 1, 6, and 9), 50 ng/ml (lanes 2,4, 7and 10) and 25ng/ml (lanes 3, 5, 8, and 11); 100 pg of eachDNAsubstratewas run as amarker(M). Thepolyacrylamide gelwasexposedtofilm for 18 h.

SV40 large T antigen clearly binds to this substrate to a greater extent than does polyomavirus large T antigen.

These data and those shown in Fig. 7 demonstrate that repeats of the pentanucleotide sequence -GAGGC- are

rec-ognizedand bound by both

polyomavirus

andSV40largeT

antigens under conditions which only allow for

specific

binding.

Binding of large T antigens of polyomavirusand SV40 to common DNA sequences. The data we obtained heretofore

showed that both large T antigenswere capable of binding with specificity to DNA fragments encompassingeach

oth-er's binding sites, and that they werecapable ofbindingto

synthetic DNA substrates containing repeats ofthe

penta-meric sequence -GAGGC-. To prove that both large T

antigens bind to the same sequences in

polyomavirus

and

SV40DNAand to prove that theimportant characteristic of this sequence is the repeat of the

pentanucleotide,

we

conducted the

experiments

outlined below. First, we

mea-suredthecapacity of variousunlabeledpolymers created by self-ligation of linkerstocompetewithlabeledpolyomavirus

DNA as substrate in binding reactions with polyomavirus large T antigen. We repeated this

experiment

with SV40

DNA and SV40 large T antigen. Second, we determined

whether thesynthetic DNAs could competeeffectivelywith

SV40 DNA as a substrate when polyomavirus large T

antigen was used as a source ofprotein and whetherthese

synthetic DNAs could act as competitors in binding

reac-tions with polyomavirus DNA and SV40 large T antigen. To conduct these experiments, we end labeled the DNA

fragments obtained by restriction endonuclease cleavage of various recombinantplasmids and reacted these with a fixed

quantity of large T antigen and various amounts of

on November 10, 2019 by guest

http://jvi.asm.org/

[image:7.612.378.497.74.315.2] [image:7.612.108.251.419.645.2]

932 POMERANTZ AND HASSELL

A

Xho

U

U1

isaa

li'.

_lip

to-__

__

M1234

B

BamHIBclI Cla I

6

U0

M 1M34M5

M 12M34 M56

FIG. 8. Immunoprecipitation of ligated synthetic linkers after

reaction with nuclearextract from polyomavirus- orSV40-infected

cells. (A) Synthetic XhoI linkers were phosphorylated and ligated

and then reactedataconcentration of25ng/ml (lanes1and3)or50

ng/ml (lanes 2 and 4) with 30 ,u1 of nuclear extract from

SV40-infected CV-1 cells (lanes 1 and 2) or polyomavirus-infected 3T6

cells(lanes 3 and 4). The marker lane (M) contained the phosphory-lated ligated XhoI linkers usedassubstrate. The polyacrylamide gel was exposed to filmn for 6h. (B) Synthetic BamHI, BclI, and ClaI

linkers were phosphorylated and ligated and then reacted at a

concentration of 25ng/ml(lanes1, 3, and 5)or50ng/ml(lanes 2, 4,

and 6) with 30 p.l ofextract from SV40-infected cells. The marker lanes (M) contain a sample of each DNA substrate used in the

binding reactions. The polyacrylamide gelwas exposed to film for 18 h.

beled, self-ligated linkers. The latter included those linkers

used previously, namely BamHI, BclI, ClaI, and XhoI. First,we measured thecapacity of self-ligated XhoI linkers,

which contain repeats of the pentanucleotide, to compete

with the end-labeled Hinfl fragments ofpPH1-8 DNA (the

polyomavirus DNA-containingplasmid; Fig. 2A) in binding

reactions withanuclearpreparation of polyomavirus large T antigen (Fig. 9A). They reveal that the self-ligated XhoI linkers servedas very effectivecompetitorswith the

natural-ly occurring polyomavirus sequences in binding reactions

with polyomavirus large T antigen. By contrast, the

self-ligated BamHI, BclI, and ClaI linkers did not compete with polyomavirus DNA for thebinding ofpolyomavirus large T antigen, evenatconcentrationsatleast 10timesgreaterthan that required for effective competition with the self-ligated XhoI linkers (Fig. 9B).

We repeated this experiment with the same competitors

but withSV40 DNA as substrate andSV40 largeT antigen

(Fig. 10A and B). The results were the same as those

reported above. The self-ligated XhoI linkers competed effectively withSV40 DNA for the binding ofSV40 large T antigen, but none of the other self-ligated linkers did. In

addition, we also observed thatXhoI linkers which had not

beenligated didnotcompetewithSV40DNAforthebinding of theSV40largeTantigen (Fig.10A). Thissameresult was

obtained when monomers of the XhoI linkers were used as a competitor inbinding reactions with polyomavirus DNA and polyomavirus large T antigen (data not shown). This result was not unexpectedbecause the -GAGGC- sequence is only formed when the XhoI linkers are ligated to each other.

We also measured the capacity of the various self-ligated linkers to compete with SV40 DNA for the binding of polyomavirus large T antigen and vice versa. The results (Fig. 11) revealed that only the self-ligated XhoI linkers competed with SV40 DNA for the binding of polyomavirus large T antigen. This is indicated by the decrease in the intensity of theAvaII-HindIll2fragment that accompanied the inclusion of increasing quantities of self-ligated XhoI linkers in the binding reaction (Fig. 11A). Similarly, only the self-ligatedXhoI linkers competed effectively for the binding ofSV40 large T antigen to the Hinfl 4 fragment ofpPH1-8 DNA (Fig. 12). The latter DNAfragment contains bothsites of binding for polyomavirus large T antigen. These results strongly suggest that both large T antigens recognize and bind to the same sequences in eitherpolyomavirus orSV40 DNA and that a crucial featureof thesesequences isarepeat of the pentanucleotide -GAGGC-.

DISCUSSION

We tested the hypothesis that repeats of the sequence -GAGGC- form the recognition-binding site for polyoma-virus andSV40large T antigens. First, we showed that both large Tantigens bind with specificity to each other'sbinding sites. A common feature of the sequences comprising the large-T-antigen-binding sites in polyomavirus and SV40 DNA is repeats of the pentanucleotide -GAGGC-. Second,

wemeasured the capacity of both large Tantigens tobind to synthetic DNA substrates containing repetitionsofthe pen-tanucleotide and to controlDNAs thatdo not andfound that only those DNAs with the pentanucleotide repeatare

specif-icallybound by the large T antigens. Finally, weshowed that among the synthetic DNAs tested, only substrates

contain-ing repeats of the pentanucleotide effectively compete with

the natural DNA substrates (polyomavirus or SV40 DNA) for the binding of the two large Tantigens to either theirown

genome or theheterologous DNA. The extentofcompetition

iscomplete at high levels ofcompetitor DNA,indicatingthat the competitor can block the specific binding ofthe large T

antigens to all of the bindingsites in both polyomavirus and SV40 DNA. These data substantiate the hypothesis that repetitions of the -GAGGC- pentanucleotide form the recog-nition-binding site for the large T antigens ofpolyomavirus

andSV40. The sequences of the various DNAstowhichthe large T antigen bind are shown in Fig. 13.

Tjian (47) first proposed that the pentanucleotide was

important for the binding of the D2 hybrid large T antigen

(23) toSV40 DNA because the protein shielded the guanine

residues in this sequence from methylation by dimethyl

sulfate. Subsequently, measurements of the capacity of authenticSV40 large T antigen to bind tomutantSV40DNA

substrates confirmed theimportanceof the pentanucleotide.

For example, a deletion of 21 bp within binding site I (between nucleotides 5207 and 5186) in the cs1085 SV40 mutant genome drastically reduces the

capacity

of SV40

large T antigen to bind to thatsite (10, 29, 45). This deletion removes a pair of -GAGGC- residues in binding site I

(Fig.

13B). Similarly, DNA of the mutantcs1088, which contains

multiple mutations in bindingsite Iwhichchangetheguanine

residues in both pentanucleotides to adenines, is also not

bound by large T antigen (29, 45). Finally, SV40 large T

antigen fails to bind tosite Iof the DNAoftheSV40mutant J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:8.612.122.244.71.321.2]

d11086 (10). The genome of this mutant is deleted by 4 bp

(between nucleotides 5192 and 5197) and this results in the removal ofone ofthe -GAGGC- repeats in binding site I (Fig. 13B).

Although theoccurrenceofrepeatsofthe pentanucleotide is sufficient to account for the binding of SV40 large T antigen to sites I and II in SV40 DNA, this sequence does notoccurin site III. The latterisalow-affinitybinding site, and SV40 large T antigen is incapable of bindingtothis site when itis separated from sites I and11(9, 10, 29, 33, 43, 46,

47). Recently, DeLucia et al. (9) have identified 13 DNA pentanucleotide sites in SV40 DNA that areprotected from methylation by SV40 large T antigen and deriveda

consen-sus sequence for them (-G/T A/G GGC-).

Large-T-antigen-binding site III contains six ofthese pentanucleotides ar-ranged as direct repeats withina 59-bp span of DNA (Fig. 13B). It is likely that repetition of these weak

recognition-binding sequences (-TAGGC-, -GGGGC-,and-TGGGG-)in

binding site IIIallows for the stablebindingofSV40large T antigen only after a fraction of the protein has already covered sites I andII.

The -GAGGC-pentanucleotide alsooccurswithin

large-T-antigen-binding sites I and II in polyomavirus DNA (Fig.

13A). Here, the sequence -A/TGAGGC- is found repeated three times inbindingsite I andtwoorthreetimesinbinding siteII,dependinguponthe strain ofpolyomavirusexamined.

Interestingly, the firstnucleotide in each repeatisseparated from thesame nucleotide of theadjacentrepeat by 10or 11 bp. This arrangement ofrepeats would place them on the sameface of the DNA helix. Theimportance of hexanucleo-tiderepeatfor thebinding of largeTantigeninpolyomavirus DNA is supported by the results we have presented here.

However, there is nodirect evidence that largeTantigen of

polyomavirus contactsthese sequences inthebindingsites.

We examined the sequences comprising the

large-T-anti-gen-binding

sites in polyomavirus DNA for sequences

ho-mologoustotheconsensusfamily derivedby DeLuciaetal. (9). This led tothe identification of three related sequences

which are located between the -A/TGAGGC- repeats in

binding

sites I and II (Fig. 13A). One such sequence,

-TGGGG-(between nucleotides 128 and 134), forms a con-tinuation of therepeats presentinsite I (Fig. 13A). The other two sequences are separated from each otherby 10bp, but

theyoccuratadistancegreaterthan 10bpfrom the repeats insites I andII(betweennucleotides 92 and 107) (Fig. 13A).

This set ofrepeats may constitute a third

large-T-antigen-binding site in polyomavirus DNA that is located between those identified previously.

Interestingly, there are also four other pentanucleotides

homologoustotheconsensussequencederived byDeLucia et al. (9), arranged as tandem pairs on each DNA strand

(betweennucleotides 5286 and 14) inourstrain of polyoma-virus. These sequences are located within the core of the

polyomavirus origin for DNA replication (31). Moreover,

theirarrangement is remarkably similarto the arrangement

oftheconsensuspentanucleotides in binding siteIIofSV40

(Fig. 13B).InSV40, binding site II comprises thecoreofthe SV40 origin for DNA replication (1). Whether, the pentanu-cleotides in the polyomavirus origin for DNA replication

compriseyetanotherlarge-T-antigen-bindingsite remainsto

bedetermined. In this regard itmaybe noteworthy that the A2 strainofpolyomavirus does notcontain three of thefour repeatsalluded to previously.

Theaffinity of SV40 large T antigen for the bindingsitesin SV40 DNA isnearly thesameasthatforthebindingsitesin polyomavirus DNA.However,polyomaviruslarge T antigen apparently binds tightertothebinding sites in polyomavirus DNA thanto those in SV40 DNA. This interpretation rests

A B

M C Xho! C BamHI L Clal M

2 _ 2

3 _,

4 _ __ 3

5-7 4m4I. I 4

8 4' * t , g

5-10.11 8

12,13 10-11

14,15

16 14-15

17-20 41

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

FIG. 9. Competition between end-labeled polyomavirus DNA fragments and ligated, unlabeled synthetic linkers for reaction with polyomavirus largeTantigen.(A)Theend-labeled DNAfragmentsobtained afterHinfldigestion ofpPH1-8DNA (1ng/ml) were mixed with various concentrations of ligated, unlabeled XhoI linkers (lane 2, 100ng/ml;lane 3, 500

ng/ml;

lane 4, 1,000ng/ml),and were reacted with 30

R1

of nuclearextractfrom polyomavirus-infected 3T6 cellsas described in the text. The marker lane (M) contains a sample of the end-labeled HinflfragmentsofpPH1-8DNA used as substrate. The control lane (C or lane 1) shows thefragmentsimmunoprecipitatedafter reaction of theend-labeled,HinflfragmentsofpPH1-8DNA with 30p.1 ofnuclear extract frompolyomavirus-infected 3T6cells in the absence of added linkers. The driedagarose gel was exposed to film for 6 h. (B) The end-labeled HinflfragmentsofpPH1-8DNA were mixed with unlabeled

li-gatedBamHIlinkers(lane 1, 100 ng/ml; lane 2, 500ng/ml;lane 3, 1,000ng/ml),BclI linkers (lane 4, 100ng/ml;lane 5, 500 ng/ml; lane 6, 1,000 ng/ml), or Clal linkers (lane 7, 100 ng/ml; lane 8, 500 ng/ml; lane 9, 1,000 ng/ml) and were reacted with 30

,u1

ofnuclear extract from polyomavirus-infected 3T6 cells. The control lane (C) shows the fragments immunoprecipitated after reaction of the end-labeled Hinfl fragmentsofpPH1-8DNA with 30p.1of nuclear extract frompolyomavirus-infected3T6 cells in the absence of added linkers. The lane labeled Mcontainsasample of end-labeled Hinfl-cleavedpPH1-8DNA used as a substrate in the binding assays. The dried agarose gel was exposed tofilm for10 h.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:9.612.155.470.432.609.2]

934 POMERANTZ AND HASSEiL

A B

XhoI Xho

ligatod minomers -

-3J-~_

2|w_*3____

6j44

6

ii * g-t 1

lM1t 3 4 M n32 456789g

FIG. 10. Competition between labeledSV40 DNA fragments and unlabeled self-ligated synthetic linkers as substrates for binding to SV4O large T antigen. (A) The fragments obtained after

AvaII-HindIIIcleavage ofpSVOl DNA were endlabeled and mixed with various concentrations of unlabeled ligated XhoI linkers before reaction with 30 ,ul of nuclear extract from SV40-infected CV-1 cells.ThepSVOl DNA fragments were present at a final concentra-tion of 1 ng/ml. In lane 1, no linkers were added, whereas lane 2 contained ligated XhoI linkers at a concentration of 100 ng/ml. Lanes 3 and 4 contained unlabeled XhoI linkers that had not been self-ligatedaFt concentrations of 100 and 500nglml,respectively. The dried agarose gel was exposed to film for 10 h. (B) The AvaIl-HindIII digestion products of pSVO1 DNA were end labeled and mixed with various concentrations of self-ligated BamHI, BclI, or

ClaI linkers before incubation with 30 p.1 of nuclear extract from SV40-infected CV-1 cells. The labeled DNA fragments were present at aconcentration of 1 ng/ml in each reaction. The concentrations of unlabeled ligated BamHI linkers (lanes 1 to 3), unlabeled self-ligated BcII linkers (lanes 4 to 6), and unlabeled self-ligated ClaI linkers (lanes 7 to 9)were 100, 500, and 1,000ng/mlwithrespect to the three lanes of each. The lane labeled M contained a portion of the AvaII-HindIII fragments of pSVO1 used as substrate in the bindingreactions. The dried agarose gel was exposed to film for 16 h (lanes 1 to 9).However, the marker lane (M) was exposed to film for 3 h. The final photograph was prepared by splicing lane M to the others (lanes 1 to 9).

on the fragile assumption that the preparations of the two

largeT antigens whichwecompared contained approximate-ly equal numbers ofactive large-T-antigen molecules. If the

difference in affinities forthetwo DNA substrates displayed

bypolyomaviruslarge T antigen is real,then thismayreflect

itsrequirement for alargerrecognition-binding sequence in

DNA. The hexanucleotide -A/TGAGGC- is repeated in the

binding sites inpolyomavirus DNA. Bycontrast, the

penta-nucleotide -GAGGC-is repeated in thebindingsitesinSV40

DNA (Fig. 13). A plausible explanation for these

observa-tions is that polyomavirus large T antigen preferentially recognizes the hexanucleotide, whereas SV40 large T

anti-gen recognizes thepentanucleotide. Thishypothesis is sup-ported by the observation that SV40 large T antigen binds more readily to the self-ligated XhoI linkers containing repeats of thepentanucleotide

tha~n

does polyomavirus large T antigen.

A single pentanucleotide or hexanucleotide is not likely the sole feature of the large-T-antigen-binding sites impor-tant for recognition and binding. These sequences occur

often in isolation

within SV40,

polyomavirus, and

pBR322

DNA, but DNAfragments whichbearthem arenot

invari-ably

bound

by large

T

antigen.

Forexample, the sequence

-GAGGC- occurs a total of 14 times in SV40 DNA and is

representedsix times within theboundariesofbindingsites I and II in SV40 DNA. Similarly, this pentanucleotide is

presentatotal of 20times inpolyomavirusDNA andoccurs seventimes between the borders ofthetwoknown

large-T-antigen-binding

sites.

Finally,

-GAGGC- occurs 15 times in

pBR322 DNA; however, in noinstancedo these sequences occurclosetoeachother(the shortest spacing betweenany twoofthese pentanucleotidesis 55bp). These observations

suggestthat thespecific bindingof SV40 and

polyomavirus

large T antigen to DNA is facilitated by repeats of the

pentanucleotide or the hexanucleotide within the

binding

sites.

The minimum numberof-GAGGC- repeats required for the efficientbindingofpolyomavirusorSV40largeT

antigen

to DNA islikelytwo. This is indicated by theobservations thatonlytworepeatsarerequiredfor the efficientbindingof

polyomavirus large T antigen to either binding site II in

polyomavirus DNA (34) ortobinding site I inSV40 DNA.

A B

M-.,...,--...-MTl

C

|

f

Cl

a 1 Am 1 vm_ _ _ __ -__

2

-Urn~ ~

2amai

3-~~~~~

2 ow4

Dtw

3

_4

1234

_123456789

FIG. 11. Competition between labeledSV4O DNA fragments and unlabeled self-ligated linkers as substrates for binding to polyoma-virus large T antigen. (A) The fragments obtained after AvaII-HindIII cleavage of pSVO1 DNA were labeled and mixed with various concentrations of unlabeled self-ligated XhoI linkers before reaction with 30 p.1 ofnuclear extract from polyomavirus-infected 3T6 cells. In each reaction the concentration of labeled DNA fragments was 2 nglml, whereas the unlabeled self-ligated XhoI linkers werepresent at a concentration of 100 (lane 2), 500 (lane 3), or 1,000nglml (lane 4). Lane 1 (C) contained no added linkers and served as apositive control. The lane marked M contained a portion of the labeled DNA fragments which se'rved as substrate in the binding reactions. The dried agarose gel was exposed to film for 10 h. (B) The AvaII-HindIII digestproducts ofp5 VOl DNA were end labeled and mixed with various concentrations of unlabeled self-ligated BamHI, BcII, or CIaI linkers before they were reacted with 30 p.1 of nuclear extract frompolyomavirus-infected 3T6 cells. The labeled DNA fragments were present in the various reactions at a concentration of 2nglml. The self-ligated, unlabeled BamHI linkers (lanes 1 to 3), the self-ligated unlabeled BclI linkers (lanes 4 to 6), and the self-ligated unlabeledCIaI linkers (lanes 7 to 9) were present atconcentrations of 100, 500, and 1,000 ng/ml with respect to the three lanes of each. The lane marked M contained aportionof the labeled DNA used as substrate in thesereactions. The dried agarose gel wasexposed to film for 6 h.

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:10.612.63.298.71.264.2] [image:10.612.331.547.326.498.2]

M

C_

Bjarn H I Bci Cla Xh1o 1I M

2 _

3 1 m

4

_Om ~~~~

qp

5 -7 5 -a 'a

12.13 F 14,15

16 J

7 19b

20

1 2 3 4 5 6 7 8 9 10 11 12

FIG. 12. Competition between labeledpolyomavirusDNA

frag-ments andunlabeled self-ligated linkersas substrates forbindingto SV40 largeT antigen. The fragments obtained after Hinflcleavage ofpPH1-8DNAwereend-labeled and mixed with various

concen-trations of unlabeled self-ligated linkers before they were reacted with 30,ul of nuclear extractfromSV40-infected CV-1 cells.Ineach reaction the labeledDNAfragmentswerepresentat aconcentration of 2 ng/ml.Lane 1(C) shows the labeledDNAfragments whichwere

immunoprecipitated after reaction of the end-labeled Hinfl

frag-mentsofpPH1-8DNAwith 30p.lof nuclearextractin the absenceof added linkers. Unlabeledself-ligated BamHI linkers wereaddedto those reactions whose outcome are shown in lanes 2 to 4 at

concentrationsof100(lane 2), 500(lane 3), and 1,000ng/ml (lane 4). Unlabeledself-ligatedBcIl linkerswereaddedtoseveralreactionsat

concentrations of 100(lane 5),500(lane 6),and1,000ng/ml (lane 7).

Similarly, unlabeled self-ligatedClaI linkerswereaddedtoseveral reactions atconcentrations of 100 (lane 8), 500 (lane 9), and 1,000 ng/ml (lane 10). Finally, unlabeled self-ligated Xhol linkers were added tothereactions whoseoutcome areillustrated inlanes 11 and 12.Heretheconcentrations oftheself-ligatedXhoIlinkerswere100 (lane 11) and 250 ng/ml (lane 12). The lane markedM contained a portion of the end-labeledHinflfragments ofpPH1-8DNAused as

substrate inthevarious reactions. The driedgelwasexposedtofilm for 16h.

Similarly, SV40 large T antigen binds efficiently to binding site II in polyomavirus DNA and to binding site I in SV40

DNA. Each of these large-T-antigen-binding sites contains

the-GAGGC-pentanucleotide repeatedonly twice (Fig. 13).

Interestingly, the smallest ligated XhoI linker species to be bound by eitherpolyomavirus or SV40 large T antigen is a pentamer. Pentamers of the XhoI linkers contain four re-peatsofthe-GAGGC- sequence per DNAstrand(Fig. 13C). Ifonly a tandem duplication ofthe -GAGGC- sequence is

required forbinding,then weshould have observed binding of the twolargeTantigenstodimersortrimersoftheligated

XhoIlinkers. One explanation ofthisdiscrepancy isthatthe

incapacity ofthe large T antigens to bind to ligated XhoI

linkerssmaller than pentamers mayreflect their requirement foraminimumlength ofDNA(>40bp) for efficientbinding. Alternatively, the weak binding-recognition sequences de-scribed by DeLucia et al. (9) neighboring the -GAGGC-repeats inbinding site II in polyomavirus DNA and binding site I in SV40 DNA may contribute to the facility with which thelarge T antigens recognize and bind to these sites.

There appears to be a correlation between the spacing of the-GAGGC- sequences in the binding sites in SV40 DNA and theaffinity of SV40 largeT antigen for these sites. For example, theaffinity ofSV40 large Tantigen is greatest for

site I where 7

bp

intervene between the

pentanucleotides

(Fig. 13B) (32, 48). However,siteII isboundpoorly by large

T antigen, and here the pentanucleotides are separated by

only 1bp (Fig. 13B). Bycontrast,the hexanucleotides which

arerepeatedin thelarge-T-antigen-bindingsites in

polyoma-virus DNA are allseparated by4or 5bp,and

polyomavirus

large T antigen binds with approximately equal affinity to both binding sites in polyomavirus DNA (Fig. 13A) (34).

This suggests that thespacingbetweentheputative recogni-tion sequences is important for the binding ofthe large T

antigens. Apparently when the distance between the first

base inadjacent pentanucleotidesis close to 10bp, then the

site is

tightly

bound

by large

T

antigen.

This

generalization

also applies to the weak recognition signals in site III of

SV40 DNA. Here the sequence -GGGGC- or -TGGGC- is

repeated five times, and the first base in each repeat is

separated

from thatsamebase in the

adjacent

repeatby8 to

11bp, dependingupon which pairsofrepeatsarecompared (Fig. 13B).

Therefore,

it seems

likely

that not only is the sequenceofthe

recognition

site

important

for the

binding

of

large T antigen but that the spacing between adjacent recognitionsites may also influence thetightnessofbinding.

Theobservationthatthe

large

T

antigens

of

polyomavirus

and SV40 bind with

specificity

to common DNA sequences

promptedustosearch theamino acidsequence of thesetwo

proteins

forhomology in the

vicinity

oftheir

origin-binding

domains. Initial attemptsto map the

origin-binding

domain

ofSV40

large

T

antigen

have ledto

conflicting

results

(5, 6,

18, 35, 37-39, 41). However,

recently

Morrison et al.

(30)

have reported that an amino-terminal

17,000-molecular-weight fragment ofSV40 large T

antigen

(amino

acids 1 to

130) is capable of

specifically binding

to SV40 DNA.

Be-causethe first 82 amino acids oflargeTantigenare shared

with small Tantigen, which does notbind to DNA, these

investigators

have

tentatively

concluded that the

origin-binding

domain ofSV40 largeT

antigen

is located between amino acids83 and 130.

Comparable

studies of

polyomavirus

largeT antigen have not been

performed.

Nonetheless, we

compared

theamino acid sequenceofSV40

large

T

antigen

withthatofpolyomavirus largeTantigenacross the

origin-binding

domain ofSV40

large

T

antigen (amino

acids 83 to

130), after

alignment

of the two sequences to maximize

homology

between them. Theextentofrelatednessbetween

the two

proteins

across the

origin-binding

domain is

23%,

whichis less than thatdisplayedovertheentirelength ofthe molecules

(36%).

However, much ofthis

homology

is clus-tered betweenamino acids124to130 inSV40

large

T

antigen

andaminoacids 278to284in

polyomavirus large

T

antigen.

Here, six ofthe sevenamino acidsareidentical between the

two T

antigens.

Moreover, fourofthe amino acids in this

region

of SV40 large T

antigen

are basic amino

acids,

whereas three in polyomavirus large T antigen are basic amino acids. Becausethetwo

large

T

antigens recognize

the

sameorvery

closely

relatedDNAsequences,it is

likely

that the

protein-DNA

contacts aremediated

by

thesidechains of

thesame orrelatedamino acidsin thetwo

proteins.

Theside

chains ofthe short stretch ofbasic amino acids described previouslyinpolyomavirusandSV40largeTantigenmay be thosewhichdirectlycontactthe-GAGGC-sequences in the viral genomes. Thissuggestioncanbe testedexperimentally.

Whatever the outcome of such experiments, it will also prove interesting to determine whether the two large T

antigens

canfunctionallysubstitute for each other in any of

the activities which are unique to large T antigen (i.e., repression ofearly transcriptionandinitiation of viral DNA

replication).

on November 10, 2019 by guest

http://jvi.asm.org/

[image:11.612.81.275.75.236.2]

936 POMERANTZ AND HASSELL J. VIROL.

A

Polyomavirus

G TC

--f 10 20 30

40]

so~

[so

70 80 90

GGGGACCCCTGGCCCGCTTA CTCTGGAGAA

AAAGAAGGA

GGCTTG

GAGG

TCCGAGG

ACTTG

TCAAAACAGG

ACTGGCGCCT CCCCTGGGGA CCGGGCGAAT GAGACCTCTT TTTCTTCTC

TCAACAICTCCAAGGTC

TCCdTTGAAC AGTTTTGTCC TGACCGCGGA

A

I

C AG

TATA

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~....

1O0 110 120 '140 I5O 160 170 180

TGiGAGGC,GCT GT$GGGGQCAC CCAAATTGAT ATAATTAAGC

CCCAACC,GCC

TCTCCCGCC

TCT

TTCAjGC CTCAjCCACCA TCATGGATAG

[image:12.612.80.545.71.511.2]

ACjCTCCGCGACA;CCCCGiGTG GGTTTAACTA TATTAATTCiG GGG*r1GEqG AGAAGGGIC AGT AGTGTGGT AGTACCTATC

...

T...

T...

AG

G_AG

AAAGT

AGA_T

G G T GT T

B

SV4O

.. ...

* 4 84 74. 64 54 44 34

TCCCGGCCCC'T AACTCdGCCC AiTCCCiGCCCC.TAACTCCiGCC CAJGTTCCiGCC CA.TTCTCCGC CCOCATGGCTG AGGGiCGGGGA TTGAGGCGGG TAGGGiCGGGG.ATTGAGG,CGG GT'CAAG.CGG GGGGTACCGAC

TATA

-. - o-5243

24 14 4 5227 5217

ACTAATTTTT TTTATTTATG CA GAGGCCGA GGC CGCCTCG GCCTCTGAGC TATTCCAGAA TGATTAAAAA AAATAAATAC GT CTCCGGCT CCGaCGGAGC CGGAGACTCG ATAAGGTCTT

5207 5177 5167 5157

GTAGTGAG GA GG TTTTTG GA0GicAGG C'TTTTGCAAA AAGCTTTGCA AAGATGGATA

CATCACTCCT CC AAAAAC CT[ GTOCC GAAAACGTTT TTCGAAACGTTTCTACCTAT

C Ligated Xho ILinkers

10 20 30 40 50 60 70

CCTCGAGGCC TCGAGGCCTC GAGGCCTCGA GGCCTCGAGGCCTCGAGGCC TCGAGGCTCC GAGGCCTCGA

GGAGCTCCGGAGCTCCGGAG CTCCGGAGCT CCGGAGCTCC GGAGCTCCGG AGCTCCGAGG CTCCGGAGCT

FIG. 13. Nucleotide sequence of DNA substrates to which polyomavirus and SV40 large T antigen bind with specificity. (A) The nucleotidesequenceofthe A2strain ofpolyomavirusthatincludesthelarge-T-antigen-bindingsites is illustrated and numbered according to Soedaetal.(42). The strain ofpolyomavirusused in thisresearch, whichwerefer toasAl, differs insequencefrom that of the A2 strain at severalpositions. The A/Tbpatnucleotide5isreplaced byaG/Cbpin theAlstrain, and the Al strain containsaninsertion of 2bpbetween nucleotides14and15.Moreover,theAl strain is deleted of10bp by comparisontothe A2 strain between nucleotides 44 and 54 (shown in the figure by the verticallinesbearingarrows). The dark lines above thesequenceshow the borders of the -TATA- box and theposition of the5' termini of the abundantearly mRNAs (denoted bythe dark line and the vertical lines bearing an arrow) (15). The initiation codon for translationoftheearlymRNAsisbracketed. Boxessurroundtherecognition-bindingsequence,-A/TGAGGC-. The dotted boxes contain the weakrecognition-bindingsequence(-TGGGC-,-TAGGC-,-GGGGC-)identifiedbyDeLuciaetal.(9). Thearrowsaboveorbelow the boxes depicttheorientationoftherecognition-bindingsequenceoneach DNAstrand.(B)The nucleotidesequencescomprisingtheSV40 large-T-antigen-binding sitesareshown. Thesymbolsusedarethesameasthosedescribed above.(C)The nucleotidesequenceofligatedXhoI linkers (-CCTCGAGG-) is shown. The symbols used are the same asthose described above. However, the boxes wereomitted from the figure becausetheywould have cluttered thedrawing. Note thatthe -GAGGC- sequence highlighted byahorizontal arrow ononeDNA strand partially overlaps with anotheron theopposite DNAstrand.

ACKNOWLEDGMENTS LITERATURE CITED

1. Bergsma, D. J., D. M. Olive, S. W. Hartzell, and K. N. We thank Monica Naujokas and Cathy Collins for excellent Subramanian.1982.Territorial limits and functionalanatomyof technical assistance and acknowledge the contributions of Ron the simian virus 40 replication origin. Proc. Natl. Acad. Sci. McKay with whom webegantheseexperimentsin 1981. U.S.A. 79:381-385.

This research wassupportedby theMedicalResearchCouncil of 2. Bradley, M. K., J. D. Griffin, and D. M. Livingston. 1982. Canada and the National Cancer Institute of Canada. J.A.H. is a Relationship of the oligomerization to enzymatic and DNA research associate of the National CancerInstitute of Canada. binding propertiesof theSV40largeT-antigen.Cell 28:125-134.

on November 10, 2019 by guest

http://jvi.asm.org/

3. Buchman, A. R., L. Burnett, and P. Berg. 1980. The SV40

nucleotide sequence, p. 799-823. In J. Tooze(ed.),DNAtumor

viruses, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring

Harbor, N.Y.

4. Carroll,R.B.,L.Hager,and R.Dulbecco. 1974.SV40T-antigen

bindstoDNA. Proc. Natl. Acad. Sci. U.S.A. 71:3754-3757.

5. Chaudry, F.,R. Harvey, and A. E. Smith. 1982. Structure and

biochemical functions of four simian virus 40 truncatedlargeT

antigens. J. Virol. 44:54-66.

6. Clark, R., K. Peden, J. M. Pipas, D. Nathans, and R. Tjian.

1983. Biochemical activities ofT-antigen proteinsencoded by

SV40A genedeletion mutants. Mol. Cell. Biol. 3:220-228.

7. Cogen, B. 1978. Virus-specific earlyRNA in 3T6 cells infected

byatsA mutant ofpolyomavirus. Virology85:222-230. 8. Cowan, K., P. Tegtmeyer, andD. D. Anthony. 1973.

Relation-shipofreplicationandtranscriptionofSV40 DNA. Proc. Natl.

Acad. Sci. U.S.A. 70:1927-1931.

9. DeLucia, A., B. A. Lawton,R. Tjian, and P. Tegtmeyer. 1983.

Topography ofSV40 A protein-DNA complexes: arrangement

ofpentanucleotide interaction sitesatthe originofreplication.

J. Virol. 46:143-150.

10. DiMaio, D.,and D. Nathans.1982. RegulatorymutantsofSV40.

Effect of mutationsataTantigenbindingsiteonDNA

replica-tion andexpression of viralgenes. J. Mol. Biol. 156:531-548.

11. Fanning,E.,C.Burger,and E. G. Gurney.1981. Comparisonof

T-antigen-associationhostphosphoproteinsfromSV40-infected

andtransformed cells of differentspecies.J.Gen. Virol.

55:367-378.

12. Fanning, E., B. Novak, and C. Burger. 1981. Detection and

characterization ofmultiple forms of simian virus 40 large T

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

13. Friedmann, T.,A.Esty,P.Laporte,and P.Deininger.1979. The

nucleotide sequence and genome organizationof the polyoma

early region: extensive nucleotide homology and amino acid

homologywith SV40. Cell 17:715-724.

14. Gaudray,P.,P.Clertant,and F.Cuzin. 1980. ATP

phosphohy-drolase(ATPase) activityofapolyomavirusT-antigen. Eur. J.

Biochem. 109:553-560.

15. Gaudray, P., D. Tyndall, R. Kamen, and F. Cuzin. 1981. The

high affinity binding site on polyomavirus DNA for the viral

largeTprotein. Nucleic Acids Res. 9:5697-5710.

16. Giacherio, D., and L. P. Hager. 1979. A poly (dT)-stimulated

ATPaseactivity associated with SV40 largeT antigen. J. Biol.

Chem. 254:8113-8116.

17. Gidoni, D., A. Scheller, B. Barnet,P. Hantzapoulos, M. Oren,

andC. Prives. 1982. Different forms of simian virus 40 largeT

antigen varyingin their affinities for DNA. J. Virol. 42:456-466. 18. Gluzman, Y.,and B. Ahrens. 1982. SV40earlymutants thatare

defectivefor viral DNAsynthesisbutcompetentfor

transforma-tion of culturedrat and simian cells. Virology 123:78-92.

19. Greenspan, D. S., and R. B. Carroll. 1981. Complex ofSV40

large T-antigen and 48,000-dalton host tumor antigen. Proc.

Natl. Acad. Sci. U.S.A. 78:105-109.

20. Griffin,B. E. 1980. Structure andgenomic organizationofSV40 andpolyoma virus, p. 61-123. In J. Tooze (ed.), DNA tumor

viruses, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring

Harbor, N. Y.

21. Griffin,J.D., G.Spangler,and D.M.Livingston. 1979. Protein

kinase activity associated with SV40 T-antigen. Proc. Natl.

Acad. Sci. U.S.A. 76:2610-2614.

22. Harlow,E.E.,D. C.Pim,and L. V.Crawford.1980.Complexof

simian virus 40 large-T antigen and host

53,000-molecular-weight protein inmonkeycells. J. Virol. 37:564-573.

23. Hassell, J., A.Lukanidin, G. Fey,andJ. Sambrook. 1978. The

structure and expressionoftwodefective adenovirus 2/simian

virus 40hybrids. J.Mol. Biol. 120:209-247.

24. Hassell, J. A., C. Mueller, A.-M. Mes, M. Featherstone, M.

Naujokas,B.Pomerantz,and W. Muller.1982. The construction ofpolyoma virus vectors: functions required forgene

expres-sion, p.71-77. In Y. Gluzman(ed.), Eukaryotic viral vectors.

ColdSpring HarborLaboratory, ColdSpring Harbor, N.Y. 25. Hassell, J. A., W. C. Topp, D. B. Rifkin, and P. E. Moreau.

1980.Transformation ofratembryofibroblastsbycloned

poly-oma virus DNA fragments containing only part of the early

region. Proc. Natl. Acad. Sci. U.S.A. 77:3978-3982.

26. Lane,D.P., and L. V.Crawford. 1979. T-antigenis boundto a host protein in SV40-transformed cells. Nature (London) 278:261-263.

27. McCormick, F., and E. E. Harlow. 1980. Association of a murine53,000-daltonphosphoproteinwith simian virus 40 large-Tantigenin transformed cells. J. Virol. 34:213-224.

28. McKay, R. D. G. 1981. Binding of a SV40 T-antigen related protein toDNA.J. Mol. Biol. 145:471-488.

29. McKay,R.,and D. DiMaio.1981.BindingofanSV40T antigen-relatedproteintothe DNAofSV40regulatory mutants.Nature (London)289:810-813.

30. Morrison, B., M. Kress, G.Khoury, and G.Jay. 1983. Simian virus 40 tumor antigen: isolation of the origin-specific DNA-bindingdomain. J. Virol. 47:106-119.

31. Muller, W. J., C. R. Mueller, A.-M. Mes, and J. A. Hassell. 1983. Thepolyomavirus originfor DNAreplicationiscomprised ofmultiple geneticelements. J. Virol. 47:586-599.

32. Myers, R. M., D. C. Rio, A. K. Robbins, and R. Tjian. 1981. SV40gene expressionis modulated bythecooperative binding ofT-antigen toDNA. Cell 25:373-384.

33. Myers, R.M.,and R.Tjian. 1980. Construction andanalysisof SV40 originsdefective in tumorbindingand DNAreplication. Proc. Natl. Acad. Sci. U.S.A. 77:6491-6495.

34. Pomerantz, B. J., C. R. Mueller, and J. A. Hassell. 1983. Polyomavirus large T antigen binds independently to multiple unique regionson the viral genome. J. Virol. 47:600-610. 35. Prives, C.,B. Barnet,A.Scheller,G.Khoury,andG.Jay.1982.

Discreteregions of simian virus 40largeTantigenare required fornonspecific and viralorigin-specific DNA binding.J. Virol. 43:73-82.

36. Prives, C., Y. Beck, G.Gidoni, M.Oren, andH. Shure. 1980. DNA-bindingand sedimentationproperties ofSV40T-antigens synthesized in vivho and in vitro. Cold Spring Harbor Symp. Quant.Biol. 44:123-130.

37. Prives, C., Y.Beck, and H. Shure. 1980. DNA-binding

proper-ties of simian virus 40 T antigens synthesized in vivo and in vitro. J. Virol. 33:689-696.

38. Rundell, K.,J. K.Collins,P.Tegtmeyer,H. L.Ozer, C.J. Lai, and D. Nathans. 1977. Identification of simian virus 40 protein A.J. Virol. 21:636-646.

39. Scheller, A., L. Covey, B.Barnet, and C. Prives. 1982. A small subclass ofSV40T antigenbinds tothe viralorigin of replica-tion. Cell 29:375-383.

40. Shalloway, D., T. Kleinberger, and D. M. Livingston. 1980. MappingofSV40DNAreplication origin region bindingsitesfor the SV40T-antigen by protection againstexonuclease III diges-tion. Cell 20:411-422.

41. Shortie, D.,R. F.Margolskee,andD.Nathans. 1979.Mutational analysisof theSV40replicon: pseudorevertantsofmutantswith a defective replication origin. Proc. Natl. Acad. Sci. U.S.A. 76:6128-6131.

42. Soeda, E., J. R. Arrand, N. Smolar, J. E. Walsh, and B. E. Griffin. 1980. Coding potential and regulatory signals of the polyomavirus genome. Nature (London)283:445-453. 43. Tegtmeyer, P., B. Anderson, S. B. Shaw, and V. G. Welson.

1981. Alternative interactions of theSV40Aproteinwith DNA. Virology 115:75-87.

44. Tegtmeyer,P.,M.Schwartz,J. K.Collins,and K.Rundell.1975. Regulationoftumor antigen synthesis by simian virus 40gene

A.J. Virol. 16:168-178.

45. Tenen,D.G.,L. L.Haines,andD. M.Livingston. 1982. Binding ofananalogof theSV40T-antigentowild-typeandmutantviral replication origins.J. Mol. Biol. 157:473-492.

46. Tjian,R. 1978. Thebindingsite onSV40 DNAfora T-antigen-relatedprotein. Cell 13:165-179.

47. Tjian, R. 1978. Protein-DNA interactionsattheorigin ofSV40 DNA replication. Cold Spring Harbor Symp. Quant. Biol. 43:655-662.

48. Tjian, R. 1981. T-antigenbindingand the control ofSV40gene

expression. Cell 26:1-2.

49. Tjian,R.,and A.Robbins. 1979.Enzymaticactivitiesassociated with a purified SV40 T-antigen-related protein. Proc. Natl. Acad. Sci. U.S.A. 76:610-614.

on November 10, 2019 by guest

http://jvi.asm.org/