Inhibition of structural changes in the simian virus 40 core origin of replication by mutation of essential origin sequences.

(1)

0022-538X/92/095248-08$02.00/0

Inhibition of Structural

Changes

in the Simian Virus 40

Core

Origin

of

Replication by Mutation of Essential

Origin

Sequences

JAMES A. BOROWIEC

Departmentof Biochemistry, New York University MedicalCenter,

New

York,

New York10016

Received 15May 1992/Accepted 16 June 1992

Mutation of the simian virus 40 (SV40) originofreplication (ori)has revealed the presence of three critical domains needed for DNAreplication.Theoutertwo

domains,

the AT tract and

early

palindrome

element (EP), colocalize with DNAregionsthat become

structurally

altered in the presence of theSV40largetumor antigen (Tantigen)and ATP. Mutations within each domainwereexamined for their effectonthe distortion ofori DNA by Tantigen,asassayed bythe

sensitivity

of DNA toKMnO4oxidation. We have found that mutations in the AT tractthat inhibitSV40DNAreplicationalso inhibit the distortion of the ATtract.

Similarly,

mutations in the EP inhibited the generation of structural changes in this element by T

antigen.

Although

AT-tract mutations or mutationson the late side of ori affected structural

changes only

in the AT

tract,

certain EP mutations ormutations onthe

early

side ofoni also inhibited AT-tract distortion. Mutation of the flanking regions did not

significantly

affect either the

affinity

of T

antigen

foroni ortherate of

binding

tooni. We conclude from these results that the primaryfunction of the

flanking

ori domains is to

undergo

structural changes required during the initiation ofSV40 DNA

replication.

Moreover,

our results

suggest

that the

efficiencyofreplicationinitiation is

significantly

affected

by

the

degree

towhich the

flanking

elementsundergo a structuraltransition.

The initiation of DNAreplication from internal sites on chromosomespresentsa

challenging

obstacle forthecellular

replication

machinery.

Sufficient

lengths

of double-stranded

DNAmustbe rendered

single

strandedto serve astemplates

for DNApolymerasestosynthesize thenascentstrands. The mannerby which

replication

factorsinteract withthe DNA

helixto convert it tosingle-stranded DNA remains poorly understood.The useofrelativelysimple modelsystemsfor eukaryoticDNA replication has made this problem tracta-ble.Examination ofsimian virus40(SV40) indicatesthat the

presynthetic changes in DNA structure take place in an elaborate series of steps(2).

An essential first step in SV40 DNA replication is the

binding of the SV40 largetumor antigen (T antigen) to the viral origin of replication (on). At elevated temperatures

(37°C),

stablebindingof Tantigento coresequenceswithin onrequiresATPandresults intheformationof amultimeric

complex containingup to12 monomersof protein (4, 14, 24).

This ATP-dependent complex is notable for its bifurcated structurewith each halfcontaining a hexamer of T antigen

(10, 24, 28). In the presence of human single-stranded

DNA-bindingprotein (human SSB, RP-A, RF-A [19, 35, 36]) or certain heterologous SSB proteins, the intrinsic DNA

helicase activity of T antigen can catalyze the large-scale

unwinding of DNA molecules fromon (8, 22, 37). Mecha-nistic parallels to these events were observed in prokaryotic model systems such as

oriC

ofEscherichia coli and

oriA

of

bacteriophage X(1, 17).

The nucleation of DNA melting in the SV40 system appears to take place within the ATP-dependent complex.

Chemical probing of the on DNA within this complex reveals two regions of altered DNA structure. In one arm of

an

imperfect

inverted repeat, termed the early palindrome

element(EP)orinvertedrepeat,approximately 8 bp of DNA

are denatured. On the opposite flank of the on situated toward the late gene side of SV40, nearly 20 bp of an adenine-thymine (AT)tractundergoaconformational distor-tion whileremaining primarily double stranded(3, 5, 26,28).

These observed structuralchangesappear tolead,atleast in part, to topological untwisting of the on-containing DNA

(11, 29).

That each of these conformationalchanges in onstructure isacritical stepduring the initiation of SV40 DNA

replica-tion is suggestedby high-resolution genetic mapping of on. Tegtmeyer and colleagues have described three discrete elements within the65-bpcoreon that are essentialforSV40 DNA replication in vivo or in vitro (7, 12, 13, 15). Twoof thesefunctional elementsclosely overlap the two conforma-tionally flexible regions inthe EP and AT-tract regions. The thirdcentral region, containing four GAGGCelements in a perfect mirror repeat, acts as a recognition element for T-antigen binding. Point mutations in any of these three elements can reduce DNAreplication activity by 2 orders of magnitude in vivo (12, 13, 15).

Therelationship between the induction of conformational changes within on and theinitiation of replication was tested byexamining the effect of replication-defective mutations on thedistortion of the two flanking regions of on by T antigen.

Single-base-pair

mutations in critical late side elements

significantly inhibited the generation of structural changes within the AT tract while having only minor effects on structural changes in the opposite flanking region. Con-versely, while all early side mutants had significantly lower levels ofstructural changes in the EP, two of these mutants were also defective in the generation of conformational changes in the AT tract. These results provide direct evi-dence that theimportance of the two flanking regions lies in

5248

on November 9, 2019 by guest

http://jvi.asm.org/

(2)

during the initiation ofSV40DNAreplication. 5230

5243M

1l l1i 31 lategenes -_

MATERIALSAND METHODS

Preparation of protein and DNA reagents. T antigen was purified from Sf9 insect cells infected with recombinant baculovirus vEV55SVT (25) or Ac941SVT. Ac941SVT, a kind gift of Monika Lusky, Cornell University Medical

College, New York, N.Y., wasprepared by initially cloning a T-antigen cDNA into the transfer vector pVL941 (23).

Ac941SVT, although producing T antigen identical in all

respects to thatisolated from

vEV55SVT,

yields three to five times more T antigen per liter of infected Sf9 cells. Prepara-tion of infected-cell lysates and immunoaffinity purification

ofTantigenweredone by the method of Borowiec et al. (3). The DNA constructs used in these experiments were a kind gift of Peter Tegtmeyer, State University of New York,

Stony Brook. The nomenclature used by Tegtmeyer and

colleaguesin reference 7 ismaintained in this paper, with the

additional change that each pointmutation is termed bs (for

base). For example, the on mutated at position 5212 is

termed bs5212. Each DNA was prepared from DHSa cells

(21) by alkali lysisandCsClbanding (30).

KMnO4 footprinting of DNA. Standard reactionmixtures

(30

RI)

contained 40 mM creatinephosphate (di-Trissalt[pH 7.8]), 7 mM MgCl2, 0.5 mM dithiothreitol, 0.5 ,ug of the

appropriateplasmidon DNA, 4 mM ATP, 25 ,g of creatine

phosphokinase per ml, andvarious quantities of T antigen.

Reactionswereincubated at 37°C for 60 min before

modifi-cationwith KMnO4. Although Tantigen can form a stable

complexwithon DNAaftershortertimes(e.g.,seeFig. 7),

the induction ofstructural changes that are detectable by

KMnO4 requires incubation times in excess of 30 min (3). KMnO4was added from a freshly prepared stock solution (200 mM) to give a final concentration of 20mM, and the reaction was incubated for an additional 4 min. Primer

extension and gel electrophoresis conditions were

essen-tiallyasdescribedby Borowiec and Hurwitz (5).

The level of KMnO4oxidationin eachregionwas

quanti-tatedby densitometric scanningofautoradiographswith an LKB Ultrascan XL laser densitometer. Autoradiographs

with various exposure times were used to ensure that the bandintensitywaswithinthe linear rangeoffilmsensitivity.

The extentof KMnO4 oxidationof theearlypalindromewas determinedby

quantitating

the

intensity

of the

region

corre-sponding to

top-strand

thymine

residues 5217 and 5218.

Although thethymine atbs5214

produces

a

KMnO4

signal,

thissignal constitutes

only

approximately

15%of the

signal

derivedfrom the

major

reactive nucleotidesat

positions

5217

and 5218 andcanbe

easily distinguished.

For the ATtract, bands

corresponding

to

thymine

residues

16, 20, 29,

and30 of the top strandwere used. The overall

intensity

of each

regionwas

separately

normalized

by

comparison

with four

control

thymine

residues located outside theSV40coreon. The mutant originswere alwaysprobed in

parallel

with the

wild-type

(wt) on,

which servedas an internal control. The maximumnormalizedvaluefor the

early

palindrome

andthe AT tract of wt ori for each titration was set at 1.0. The

fraction of maximum oxidation for the mutant

origins

was then calculatedforeach datum

point.

Gel

mobility

shift

analysis

of T

antigen-oi complexes.

The

32P-end-labeledori DNAwas

prepared by

first

digesting

the

appropriatemutant

plasmid

DNAwith EcoRI andHindIIIto release the

oni-containing

DNA

fragment

(approximately

100 bp in

length).

After treatment with calf intestinal alkaline

> (18,41)

L4bslS

34)

--(8,31) bs521 ' j(6,27) bs26

bs5214 Ins5243 (1,22)

(8,16) (<1, 5)

Inv Pent1 InvPent 4 (<1, 1) (<1, 1)

FIG. 1. Mutations in theSV40 coreonthatwere tested for their effect on the distortion ofon DNA by T antigen. The general location of eachmutant is shown. bs5211, bs5212, and bs5214 are located in the EP; Inv Pent 1,Ins5243, bs4, and Inv Pent 4 are located in the central GAGGCelement (GAGGC); andbsl8,bs22, and bs26 are located in the AT tract. The numbers below each mutation indicate the replication efficiency relative to wt on (in percentages). The first number indicates the activity in vivo, while the second number indicates the activity in vitro, obtained by using crude cytoplasmic extracts of HeLa cells (data from reference 7).

phosphatase (USBiochemical), the DNA was end labeled in thepresence of

[_y-32P]ATP

andpolynucleotide kinase, and the labeled on DNA was purified after electrophoresis through an 8% polyacrylamide gel (acrylamide/bisacryl-amide ratio, 29:1). Each gel shiftassay used approximately 50 fmol ofDNA(at 1 x

106

to2 x

106

cpm/pmol). For gel

mobility shift assays, standard reaction mixtures (20 ,ul)

contained 40 mMcreatinephosphate (di-Tris salt[pH7.8]),

7 mM MgCl2, 0.5 mM dithiothreitol, 4 mM ATP, 25

F±g

of creatinephosphokinaseperml,onDNA, and various

quan-tities of T antigen. Standard binding reactions were incu-bated at 37°C for 60 min and then cross-linked by the

additionofglutaraldehydeto0.1%andincubationfor 15 min at 37°C. Kinetic measurements used shorter binding times

(see Fig. 8), followed by the identical cross-linking proce-dure. The reaction mixtures were then loaded onto a 4% polyacrylamide gel (acrylamide/bisacrylamide ratio, 29:1) andelectrophoresed in thepresence of 25 mMTris, 190 mM

glycine,and 1 mM EDTA(pH 8.5). Gelslicescorresponding

to the shiftedT-antigen-DNAcomplexwere removed, and theradioactivitywasdeterminedbyscintillation counting.

RESULTS

Thebindingof Tantigen tothe SV40

on,

inthepresence of ATP, results in the induction of

significant

structural changes at two flanking regions within on. Each of these deformable DNA elements resides in

regions

found

previ-ouslytobecritical foronfunction. Wethereforewishedto understandthe

relationship

between thestructural deforma-tion of these elements

by

T

antigen

and the

importance

of these elements for DNA

replication.

We used a series of mutations located within on thatwere

initially

constructed

by Tegtmeyer andcolleagues

(7,

12, 13,

15).

Themutations werechanges of one to three base

pairs

or

single-base-pair

insertions andwere located withinone of the three critical

regionsof

on,

i.e.,the

flanking

EPorAT-tract

regions

orthe central GAGGC element

(Fig. 1).

Each mutant on was

presentas a coreon

lacking

the

flanking

T-antigen-binding

site Iand SV40

transcriptional

control elements.

The induction ofstructural

changes

in each mutant on,

contained in

plasmid DNA,

was

probed by using KMnO4

oxidation. While

KMnO4

reacts

poorly

with DNA in the

-.

I

on November 9, 2019 by guest

http://jvi.asm.org/

[image:2.612.321.560.81.198.2]

(3)

A.

OriginDNA wtori bs22

I I

Tantigen(gLg) 0.0 0.2 0.4 0.8 0.0

.. 4.

.... j,:44

' I

0.2 0.4 0.8

. 'S

40 g

4 _m

B.

OriginDNA wtori bs5212

Tantigenhig) 0.0 0.2 0.4 0.8 0.0 0.2 0.4 0.8 P.

* ; ^ ^

J5214

jji _ 521

V * -521e

w w# -5225

EP

GAGGC

-4

a a

- _16

- 9 -20 AT

_29

* * * * -30

_ t_ * I e v

. S It

. _!

a. u l

.q

-m...

** 5214

* _5217

* .-5218 -5225

-_

.2!l _ 16

U,.~~~~~~~~~~~~~~~~.

-zo

L.*X.t

ft-3g~_2

12 3 4 5 6 7 8 1 2 3 45 6 7 8

FIG. 2. Effect ofT-antigenlevelsonthe induction of structuralchangesin thewild-typeonandinonDNA moleculesmutated in theAT

tractorEP. Arepresentative autoradiograph showingthechangesinDNAreactivitytoKMnO4as afunctionof Tantigenis shown. Towt

on(AandB;lanes 1to4),bs22(A;lanes5to8),orbs5212(B;lanes5to8)wasadded 0.0,ug (lanes1 and5),0.2jig(lanes2 and6),0.4,ug

(lanes3 and7),or0.8pug (lanes4 and8)of Tantigen.After incubation for60min,the DNAwasmodified with 20 mMKMnO4.Theoxidation reactionwasthenquenchedand the sites of oxidationweredeterminedasdescribed in Materialsand Methods. Varioussequencepositions

areindicatedontherightsideof eachpanel.The locations of theEP,centralGAGGC,and ATtractareshownonthe leftside of eachpanel.

B-form, DNA that is meltedor apparently bent sharplyor

significantly untwisted (of altered helical pitch) becomes hyperreactive to KMnO4 oxidation, predominantly at thy-mine residues(5, 6).These residuescanbepreciselylocated by first extending a 32P-labeled DNA primer across the region of interestwith the Klenow fragment of DNA

poly-merase I. Because the oxidized nucleotides can stall the polymerase, analysis of the extensionproducts by denatur-ing gel electrophoresis and autoradiography indicates the sites and relative extentofKMnO4oxidation. With asingle

exception (bs5214; seeMaterials andMethods), noneof the

mutations affected thymines producing the KMnO4 oxida-tionsignal.

The inductionofstructuralchangesinthewtonDNAwas

first compared with an on DNA mutated in the AT-tract

element at nucleotide 22 (bs22). The bs22 mutation was

foundpreviouslytoinhibit SV40DNAreplicationto

approx-imately 40% ofwt levels when tested in vitro with crude extracts,and under 20% ofwtlevels whenexamined in vivo (7, 13, 15). These DNA molecules were incubated with

increasing levels of T antigen, and the DNA was then

modified with KMnO4. A representative autoradiograph of theprimer extension productscomparesthe KMnO4

oxida-tionpatternsforwtonand bs22 (Fig.2A). As theamountof Tantigen addedtowtonwasincreasedfrom 0.0to0.8 ,ug, thelevel of KMnO4oxidationalso increasedat tworegions withinon(Fig. 2A, lanes1to4).These regions,asindicated

onthe left side ofthe figure,correspond to the EP andAT tractofon.Titrationof similaramountsof T antigentobs22 induced identical nucleotides withinthis mutant tobecome hypersensitiveto KMnO4 oxidation (Fig. 2A, lanes5 to8).

Visual comparison of the amount of KMnO4 oxidation in thesetwoon DNAmolecules indicatedthatmodificationof the ATtract of bs22was significantlydecreased relative to wt on. Although oxidation of the EP element of bs22 appeared slightly less than that of wt on, quantitative

analysisof the data foundnosignificantdifference(see Fig. 3B). Thus,in this qualitative comparison, asinglebase pair mutation of the AT tract inhibited thegeneration of struc-turalchanges in the ATtractbyTantigen.

A similar experiment was performed comparing the

KMnO4oxidation patternsofwtonandan onDNAmutated

in the EP at nucleotide 5212 (bs5212; Fig. 2B). This point mutation inhibits SV40 replication tolevelsthat are31% in

vitro and8%in vivoofwtlevels (7, 12, 15). The EP mutation significantly decreased the extent ofKMnO4 oxidation in thisregionwhenintermediate levelsofTantigenwereused

(0.2and0.4 pg;comparelanes 2and 3 with lanes 6 and 7 in

Fig. 2B). Noobservableeffects of thebs5212 mutationwere

notedonmodificationoftheATtract.Thus, mutations in the EPorATtractthatareinhibitorytoSV40 DNA replication also inhibitdistortion ofthe mutated region by T antigen.

We performed a comprehensive analysis of the effectof on mutationon thedistortion ofon structureby T antigen. Eachmutation testedwas foundpreviouslyto significantly reduce SV40 replication both in vivo and in vitro (Fig. 1). Each ori mutant was incubated with various amounts of T antigen, in the presence of ATP, and the DNA was then

probedwithKMnO4. Theextentof KMnO4 oxidationfor the AT-tract and EPregionsineachon were separately

quanti-tatedbydensitometricscanning of theautoradiographs. The levels ofmodification in these twoon elementswere

sepa-EP

GAGGC

AT

on November 9, 2019 by guest

http://jvi.asm.org/

[image:3.612.131.507.80.342.2]

(4)

1.0

ATOxidation

o0.8

Cu~~~~~~~~~~~~~~~~~~~~~~~4

~0 0.6

-C

0.4-E 0.2 -0 z

0.0

0.0 0.2 0.4 0.6 0.8

[image:4.612.90.278.76.403.2]

TAntigen,gg

FIG. 4. Effect of T-antigen levelsonKMnO4 oxidation in theEP

and AT tract ofon DNA molecules mutated on the early side

(nucleotides 5211to5243). Quantitation ofoxidation in the EP (A) and ATtract(B) is shown. ThemutantonDNAs testedwerewton

(0),bs5211(D),bs5212(A), bs5214 (0),Ins5243 (A), andInv Pent 1(a). Quantitation of oxidationwasasdescribed in the legendto

Fig.3.

mutationreduced KMnO4 oxidation of the ATtracttolevels 18 to 40% of that of wton. High levels of T antigen (0.8 ,ug)

did not overcome the effect of the mutation, as KMnO4

modification within the ATtractremained 39 to66% ofwt levels. Incontrast,these identicalmutations had less signif-icant effectson the generationofstructural changes within

the EP(Fig. 3B).Inthepresenceof 0.4,ugof Tantigen,four of the five mutations reducedKMnO4modificationof the EP to levels 72 to 99% ofwt values. At 0.8 ,ugof T antigen, oxidation within the EP ofanyof thesemutantsdid notfall below 79% of the wt level. Therefore, late-side mutations that are deleterious to SV40 DNAreplication significantly inhibited AT-tract distortionbyTantigenwhilehaving only minor effectsondistortion of the EP.

Fiveearly-sideonmutantsweretested for theirabilityto

undergo T-antigen-mediated changes in DNA structure. Three of these mutations were located in the EP (bs5211,

bs5212, andbs5214).The fourth mutation was aninversion

of themostleftward(early) GAGGCsequence(InvPent 1), and the lastwas asingle-base insertion betweenthe middle two GAGGC pentamers (Ins5243). The Ins5243 mutation, because of itspositionin the center ofon,wouldpresumably separatethe two hexameric lobes ofT antigen by0.34 nm

and rotate one lobe with respect to the other by

approxi-C

~0

x

o

'a

.R

0

0.4 TAntigen, igg

0.8

0.0 0.2 0.4 0.6 0.8

TAntigen, igg

FIG. 3. Effect of T-antigen levelsonKMnO4 oxidation in the EP and AT tract of on DNA molecules mutated on the late side (nucleotides 1to31).Quantitation of oxidation in theATtract(A)

and EP(B) is shown. ThemutantonDNAs testedwerewton(0),

bsl8(-),bs22(A), bs26 (0),bs4 (A), and InvPent 4(El).Various amounts ofT antigen were incubated with the wt or mutant on

plasmid, and the sites of oxidationwerevisualizedasdescribed in thelegendtoFig. 2. The levels of KMnO4 oxidation for the EP and

ATtractwereindividuallydetermined for eachonbydensitometric

analysis of the autoradiograph. The data for each element were normalized with respectto themaximum level ofoxidation

deter-mined forwton(at0.8 jigofTantigen) by comparison with control

bands located outsideon.

ratelynormalizedby comparisonwith controlbands located outside the on region. The normalized levels of KMnO4

oxidation in the AT-tract and EPregionswere thenplotted

as a function of added T antigen for eachmutant on. The datawere grouped accordingtowhether the mutationwas

locatedon the late side (nucleotides 1 to26; Fig. 3) or the earlyside (nucleotides 5211 to 5243; Fig. 4)ofon. Control experiments examining on mutations that did not signifi-cantly affect SV40 DNA replication were also performed. These mutations (atnucleotides5222, 5228, and12)did not significantly affect the distortion of on structure (data not

shown).

Five late-side mutations were examined, and each had

severeeffectsonthe distortion of theAT tractbyTantigen

(Fig. 3A). Twowere located inthe central GAGGCregion (bs4andInvPent4),and the otherthreewerelocated within

theAT tract(bsl8, bs22,andbs26).TheInvPent 4 mutation

is an inversion of the most rightward (late) GAGGC

se-quence. At moderate levels of T antigen (0.4 ,ug), each

0₀ x

0 0

c

0

z

B

C

Y.

x

o

0

z

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.353.541.81.411.2]

(5)

OrgginrDNA vtIorl !nvPeflnt bs52 1 2

- ----r---

--Tantigen (LP, 0001 02025.4 r0 220. .2

1 2 3 4 5 6 7 8 9

11

Ii 12

FIG. 5. Effect of on mutation on the formation of the ATP-dependentcomplexbyTantigen.Variousamountsof Tantigen,as

indicated, were incubated with 32P-end-labeled DNA fragments containing eitherwton (lanes1to4), InvPent1 (lanes5 to8),or

bs5212(lanes 9 to 12), in thepresence of ATP. After 60min, the T-antigen-DNA complexes were cross-linked by the addition of glutaraldehyde. Thecomplexeswere then separated from freeon DNAbyelectrophoresis through anondenaturing gel, and thegel was dried and autoradiographed. TheT-antigen-DNA complexis indicatedbyabracketonthe left sideof thegel.

mately 34°. Each mutation

significantly

inhibited the induc-tion of structural

changes

within the EP

(Fig.

4A).

The

mutations locatedintheEP

region

reduced themodification

ofthis region 28 to 65% compared with thewton at

high

levelsof Tantigen(0.8

,ug).

The Inv Pent 1 mutation had the greatesteffecton thegenerationofstructuralchanges, with theextentofKMnO4 modification reducedtoless than2%of wtlevels.

Early-side mutationshad differentialeffects onthe induc-tion of structural changes inthe ATtract (Fig. 4B). At the

highestlevel ofT antigen (0.8 ,ug),AT-tract distortion was significantly inhibitedfortwomutants,bs5214and Inv Pent 1,to 28 and39%, respectively, ofwtlevels. The Inv Pent 1

mutationisparticularly informativesincethecorresponding mutationonthe lateside,InvPent4, hadinsignificanteffects on KMnO4 modification ofthe EP. Two other early-side

mutations

(bs5211

and

bs5212)

had only minor effects on

KMnO4 modificationof the ATtract.The last mutant tested,

Ins5243, had moderately reduced levels of AT-tract oxida-tion compared with the wton. Weconclude that although

late-sidemutations have little effect on structural changes in theEP, certain early-side mutations can affect the generation ofstructural changes in both the EP and AT tract.

Theinhibitory effect ofonmutation onondistortion could be either an indirect result of the mutation decreasing

T-antigen bindingor adirect effect of the mutation

decreas-ing the intrinsic ability of the flanking

on

elements to

undergo a conformational change. To distinguish between these twopossibilities,weused a gel shift assay to examine the binding of T antigen to a

32P-labeled

DNA fragment

containing either a wt or mutant

on.

Increasing levels of T

antigen

were added to the on DNA molecules, and the

4000

0..

E

0

~o

C

0

z a

0.0 0.1 0.2

TAntigen,

9g

0.3 0.4

FIG. 6. Quantitationof the effect ofon mutationonthe forma-tion oftheATP-dependentcomplex by T antigen.Increasinglevels of Tantigenwereincubated with 32P-end-labeled DNA fragments containingwt or a mutanton. The T-antigen-on complexeswere thencross-linkedby usingglutaraldehyde and separated fromfree on DNAbynondenaturing gel electrophoresis asshown inFig. 5. The portions of the gel corresponding to the T-antigen-ori com-plexes,afterdetectionbyautoradiography,werethenremoved, and theradioactivitypresent wasdeterminedby scintillation counting. Theon DNAconstructstestedwere wton(0), bs5212(U),bs5214 (A),InvPent1(0), Ins5243(O),bs4(A),InvPent 4

(+),

bsl8(El), bs22(x),andbs26(*).

T-antigen-DNA complexes were then cross-linked with glu-taraldehyde. TheT-antigen-DNA complexeswereseparated by nondenaturing gel electrophoresis and detected by auto-radiography.

Thebinding of Tantigen to wt on, Inv Pent 1, andbs5212

was examined first(Fig.5). The addition ofasmall amount of T antigen (0.1 jig) to wt on led to the production of two distinct complexes (lane 2). Comparison with the work of Parsons et al. (28) indicates that the faster and slower complexesareoni molecules boundbyahexamerordouble hexamer of Tantigen,respectively. Increasingthe levelof T antigen to0.4 jig increased the amountof DNAboundbya double hexamer of Tantigen (lane 4). Qualitativeinspection of theautoradiographshowsthat the Inv Pent 1mutation,as

expected,decreased theabilityof Tantigentobindtotheon (lanes 5 to 8), whereas the bs5212 mutation had little observable effect (lanes 9 to 12).

Similar conditionswereusedtoexamine thebindingof T antigentothemutantonDNAs. Thebindingof Tantigento each DNAwasquantifiedbydeterminingthe sum of radio-activity present in both bands of the complex (Fig. 6). Certain mutations located in the central GAGGC region significantly reduced thebindingof T antigen. At 0.4jigof T

antigen, the InvPent 1,bs4,and Inv Pent 4 mutations each decreased the amountofT-antigen-ori complex to

approxi-mately two-thirds of the wt level. On the other

hand,

T-antigen binding to DNA molecules containing mutations in the ATtract or EP(bs5212,bs5214,bsl8, bs22, and bs26) or to the Ins5243 mutant was not significantly different from

bindingto wton, and any small differences were within the normalvariation of this assay. Thus, mutations in the EP and ATtractdidnotinhibit the induction of the conformational

changesin theseregions by preventing T-antigen binding to on. Ourresults suggest instead that these mutations directly affect the abilityof the AT-tract and EP regions to undergo structural changes mediated by T antigen.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:5.612.339.545.75.245.2] [image:5.612.96.269.77.293.2]

(6)

I

Incubation(min) 0 0.5

---I

1 2 5 15 600.5 1 2 5 15 60

.1611

!1|111

E 2000

CL 0

m

z 1000

0 5 10 15

1 2 3 4 5 6 7 8 9 10 11 1213 FIG. 7. Effect ofon mutation on the rate of formation of the ATP-dependent complex. T antigen (0.2 ,ug) was incubated with 32P-end-labeledDNAfragments containing either wton(lanes 1 to 7)orbs5212(lanes 8 to 13). Aftervariousintervals (as indicated), the T-antigen-DNA complexes were cross-linked by the addition of glutaraldehyde. The complexes were then separated from freeon DNAby electrophoresis through a nondenaturing gel, and the gel was dried andautoradiographed. The T-antigen-DNA complex is indicatedby a bracket on the left side ofthe gel.

on mutation may in turn affect the rate of T-antigen

binding toon. To examine this possibility, we incubated T

antigen(0.2 ,ug)forvarioustimes with32P-end-labeled DNA

fragments containingamutant orwton. TheT-antigen-on complexes were cross-linked with glutaraldehyde, and the

relative amount ofcomplexformed was examined by a gel

shiftassay. AtimecourseofT-antigen bindingtothe wton

andtheEP mutantbs5212 indicated thatthe rates ofbinding tothewton (lanes 1 to7) and bs5212 (lanes 8 to 13)were verysimilar(Fig. 7).The amount of Tantigen boundtothese andother mutantoriginsateachtime pointwasquantitated

by determining the amount of radioactivity in gel slices

corresponding

to both bands of the T-antigen-DNA com-plex.Tantigen bound allmutanton DNAmoleculesat rates

similartothatforbindingto wton DNA

(Fig. 8).

Although

theoverall amountof

T-antigen

bindingto DNA

fragments

containing mutations intheGAGGCelement of on

(bs4,

Inv Pent1, and Inv Pent

4)

was reduced, the rateof

T-antigen

binding

tothese

fragments

wassimilartothat for

fragments

containingwton.

DISCUSSION

The

ATP-dependent

binding

of T

antigen

to the SV40 originofreplicationresults in the structural distortion oftwo flanking

regions

of on. These two

regions

overlap

two on

elements

required

for

origin

activity.

Wehave examined the effect of

replication-defective

onmutationsonthe distortion ofonDNAbyT

antigen.

Mutation of critical nucleotides in

either the EP or the AT tract decreased the

ability

of T

antigen to induce structural

changes

within that

region.

Theseflankingmutations haveno

significant

effectoneither the amount of T

antigen

bound or the rate of

T-antigen

bindingtoon. We thereforeconcludethat the

importance

of

Time of Binding (min)

FIG. 8. Quantitation of the effectofonmutationonthe kinetics

offormation of the ATP-dependentT-antigen-on complex. T

anti-gen(0.2 ,ug)wasincubated withwtonor amutantonfor various

times. After cross-linking of the T-antigen-DNA complex, the

complex was separated from free on DNA by electrophoresis

throughanondenaturinggel (as shown in Fig. 7). The portionsof the

gel correspondingtothe T-antigen-ori complexes,after detection by

autoradiography,wereremoved, and the radioactivitypresentwas

determined by scintillationcounting. TheonDNAconstructstested

werewton(0), bs5212(U), bs5214 (A), Inv Pent1(0),Ins5243(l),

bs4(A),InvPent4(+), bsl8(E0),bs22 (x), and bs26 (*).

the two flanking regions lies in their ability to undergo structural changes required during the initiation of SV40 DNAreplication.

Of the 10mutations tested, 6werelocatedin the flanking

sequence elements. We found similar qualitative effects of

themutationson the generation of structural changes in the

mutated region andon SV40 DNA replication. The bs5214

and bs26 mutations had the greatest effect on SV40 DNA

replication in vitro andwerealso the mostinhibitory tothe generation of structural changes in the EP and AT tract, respectively. These results suggest that the efficiency of replication initiation issignificantly affected by the degreeto

which the flanking elements undergoastructuraltransition.

Mutation of the central GAGGC element inhibited the ability ofT antigen to bind toon. Thisresult is consistent

with various studies demonstrating close contacts of T antigen with 5'-GAGGC-3' sequences in both the core on

and theadjacent site1(5, 16, 32) andagreeswith the results

of Parsons et al. (26). Somewhat surprisingly, mutation of

the core sequences inhibited only the level of T antigen

bindingtotheonbut didnotsignificantlyinhibit theoverall rate of T-antigen-on complex formation. In other words, normalization of each set ofbinding data to the maximum amount of DNA bound indicated that the binding of the mutantonDNAwasneverless than 68% of thewton level ateach time point tested. Although the GAGGC elements

are presumably requiredto properly positionT antigen on on, the rate-limiting step of binding, under our reaction conditions, does not appear to be the recognition of these

sequences. Wehavenotperformedextensive kinetic

analy-sis of thebindingof Tantigen towt on, but it wouldseem

plausiblethata "slow" stepoccursafterthe initial interac-tionofTantigenwith the DNA. Dean et al. (9)haveshown

thatpreformedhexamers ofTantigencannotproperlybind to on and that therefore T-antigen hexamers are probably

formedontheDNAbysequentialaddition ofmonomers(see

60

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.86.268.75.283.2] [image:6.612.333.541.76.249.2]

(7)

also reference 28). Apossible interpretation ofour data is that theformation of the hexameric Tantigenonon afterthe initial GAGGC recognition is the rate-limiting step in the formationof theATP-dependent complex. The useofother assays(e.g., dimethyl sulfatefootprinting)mayallow

detec-tion of intermediate complexes during the initial stages of binding.

Why does mutation of the flanking regions inhibit their

distortion by T antigen? The minimal effect of such

muta-tions on T-antigen binding demonstrates that the reduced

structural distortion is notaconsequenceofpoorT-antigen binding to the origin. However, the workbyParsons et al.

indicates that Tantigencanweakly recognizethe EPregion

in the absence of the central GAGGC and AT-tract regions (26). Thus,onepossibilityis that Tantigenstill bindswithwt efficiency to the mutant templates because of the intact

GAGGC elements, yet forms weaker contacts with the mutatedflanks. Tantigenwould therefore be unable toform

contacts sufficiently stable to fully distort the flanking

se-quence.Analternativemodel is that eachflankingelementis predisposed to undergo a structural transition and that

mutation lessens their ability to become distorted. In this

model, Tantigen can interactwith the flanking regionsbut

these contactsareinefficientlyconverted intoanobservable structural change. These two models are not mutually

ex-clusive, andour data cannotyet distinguishbetween them.

Our data bearuponthe interaction ofT-antigenmolecules

bound toon in thepresenceof ATP. Mutation of eitherthe EP or the AT tract affected primarily the generation of

structural changes within the mutant element whilehaving lesssignificanteffectsontheopposite flanking region. Thus,

these datasupport the hypothesis that the ATP-dependent complex is composed of two somewhat independent do-mains,each containingahexamer ofTantigen (3, 28). It is importanttonote that certainearly side mutations(bs5214, Inv Pent 1) significantly inhibited conformational changes

within the AT tract. Similarly, the Ins5243 mutation, a

single-base-pairinsertionbetween theearly and latehalves

ofon, reduced thelevel of structuralchangesinboth flanks

ofon. Our datatherefore suggestthat the twohexamers do

havesignificantinteractions andthat the hexamerboundto

theearlyhalf ofonappearstobedominant to the Tantigen

bound to the late half. This hypothesis has also been independently suggested by Parsons and Tegtmeyer (27),

whoexamined thespacing requirementsbetweenconserved

elements inon.

The events occurring prior to the synthesis of nascent

DNA duringthe initiation of SV40 DNA replication have beenthe focus ofsignificant effort (for areview, see

refer-ence 2). These studies suggest that T antigen, binding as

monomericunits toon, forms ahexamer over each half of on (24, 28). In the early half of on, the EP acts as the

nucleation site forDNAmeltingwithintheon (5, 26).Both

human SSB protein (RF-A, RP-A) and some noncognate SSB proteins can allow the unwinding of on-containing DNAmolecules by Tantigen (8, 22, 37). Because specific interactionsbetween T antigen and certain nonhuman SSB proteins (e.g.,E. coli SSB)seemunlikely,the major role for

SSB is presumably to bind to the single-stranded DNA

within the EP and thus further destabilize the

double-strandedstructure ofon. Wenote, however, thatT antigen canspecificallyinteract withhumanSSB (18)and that these

contacts mayfacilitatethisevent. TheATtract, relativeto

the EP, does not appear to be greatly melted within the ATP-dependent complex (5, 26). The structural changes

within theATtractwouldbothfurtherdecreasethe stability

of the double helixand allow release of theT-antigen DNA helicaseonbinding of SSB to the EP. The T antigen, inan as yetunidentified step,would release from on astwo hexam-ers,orpossiblyas adoublehexamer(33), with each T-anti-gen hexamerbinding to opposite strands so that T antigen can move outwardfrom on in the 3'-*5' direction (20, 31,

34). Each hexamer then individually unwinds the DNA primarily by usingcontactswiththe sugar-phosphate back-bone(31).

Furtherworkisnecessarytounderstandhow Tantigen is able to convertfromasequence-specific DNA-binding pro-tein to a sequence-independent DNA helicase. Our current understanding indicates that theprocess isan intricateand subtle interactionbetweenTantigenand theon, leading to the complete denaturation of the origin structure, which allows the subsequentsynthesisof nascentDNA.

ACKNOWLEDGMENTS

We thank PeterTegtmeyer for his kindgiftof the SV40 origin constructs,MonikaLuskyfor thegenerousgiftofAc941SVT,and WarrenJelinek foruseof the LKB Ultrascan XL laser

densitome-ter. We also thank Frank Dean and Jerry Hurwitz for critical

commentsonthemanuscriptand DhrubaSenGupta,Len Blackwell, and Tom Gillette forhelpfuldiscussions. Weappreciatethe techni-cal assistanceprovided byRichard Tomko and Max Kallet.

Researchwassupported byNIHgrantAI29963,the Pew Biomed-ical ScholarsProgram (T88-00457-063), andKaplan Cancer Center Developmental Funding and Cancer Center Support core grant (fromNCIP30CA16087).

REFERENCES

1. Baker, T. A., K. Sekimizu, B. E. Funnell, andA. Kornberg. 1986. Extensive unwinding of the plasmid template during staged enzymaticinitiation of DNAreplicationfrom theorigin of the Escherichia coli chromosome. Cell 45:53-64.

2. Borowiec, J. A.,F. B.Dean,P. A.Bullock,andJ.Hurwitz.1990. Bindingandunwinding-howTantigenengagesthe SV40origin ofDNAreplication.Cell 60:181-184.

3. Borowiec, J. A.,F. B.Dean,andJ.Hurwitz.1991. Differential induction ofstructuralchanges in the simian virus40originof replication byTantigen. J.Virol.65:1228-1235.

4. Borowiec, J. A., and J. Hurwitz. 1988. ATP stimulates the binding ofsimian virus 40 (SV40) large tumor antigen to the SV40originofreplication. Proc. Natl.Acad. Sci. USA 85:64-68.

5. Borowiec, J. A., andJ. Hurwitz. 1988. Localized melting and structural changes in the SV40 origin of DNA replication inducedbyTantigen. EMBOJ.7:3149-3158.

6. Borowiec,J. A., L. Zhang, S. Sasse-Dwight, and J. D.Gralla. 1987. DNA supercoilingpromotes formation ofa bent repres-sionloopinlacDNA.J. Mol. Biol. 196:101-111.

7. Dean,F., J. Borowiec,Y.Ishimi,S. Deb, P. Tegtmeyer, and J. Hurwitz. 1987. Simian virus 40 large tumor antigen requires threecore replication origin domainsfor DNA unwinding and replicationinvitro.Proc. Natl. Acad. Sci. USA84:8267-8271. 8. Dean,F.,P.Bullock,Y.Murakami, C. R. Wobbe, L. Weissbach, andJ.Hurwitz. 1987.Simian virus40(SV40)DNA replication: SV40largeTantigenunwindsDNAcontaining the SV40 origin ofreplication.Proc. Natl. Acad. Sci. USA 84:16-20.

9. Dean,F.B., J.A.Borowiec, T. Eki, and J. Hurwitz. 1992. The simian virus40 Tantigen doublehexamer assembles around the DNAatthereplication origin.J. Biol. Chem. 267:14129-14137. 10. Dean, F. B., M. Dodson, H. Echols, and J. Hurwitz. 1987. ATP-dependentformationof a specialized nucleoprotein

struc-turebysimian virus40(SV40) largetumor antigen at the SV40 replicationorigin.Proc.NatI.Acad. Sci. USA84:8981-8985. 11. Dean, F. B., and J. Hurwitz. 1991. Simian virus 40 large T

antigenuntwistsDNAattheorigin of DNA replication. J. Biol. Chem. 266:5062-5071.

12. Deb, S.,A. L.DeLucia,C.-P. Baur, A. Koff, and P. Tegtmeyer.

on November 9, 2019 by guest

http://jvi.asm.org/

(8)

13. Deb, S., A. L. DeLucia, A.Koff, S. Tsui, and P. Tegtmeyer. 1986. The adenine-thymine domain of the simian virus 40 core origin directs DNAbending and coordinately regulates DNA replica-tion. Mol. Cell. Biol. 6:4578-4584.

14. Deb, S., and P. Tegtmeyer. 1987. ATP enhances the binding of simian virus 40 large T antigen to the origin of replication. J. Virol. 61:3649-3654.

15. Deb, S., S. Tsui, A. Koff, A. DeLucia, R. Parsons, and P. Tegtmeyer. 1987. The T-antigen-binding domain of the simian virus 40 core origin of replication. J. Virol. 61:2143-2149. 16. DeLucia, A. L., B. A.Lewton, R. Tjian, and P. Tegtmeyer. 1983.

Topography of simian virus 40 A protein-DNA complexes: arrangement of pentanucleotide interaction sites at the origin of replication. J. Virol. 46:143-150.

17. Dodson, M., H. Echols, S. Wickner, C. Alfano, K. Mensa-Wilmot, B. Gomes, J. LeBowitz, J. D. Roberts, and R. Mc-Macken.1986. Specialized nucleoprotein structures at the origin ofreplication of bacteriophage X: localized unwinding of duplex DNAby a six-protein reaction. Proc. Natl. Acad. Sci. USA 83:7638-7642.

18. Dornreiter, I., L.F.Erdile,I.U.Gilbert,D. vonWinkler,T.J. Kelly, and E.Fanning. 1992. Interaction ofDNA polymerase alpha-primase with cellularreplication protein A andSV40 T antigen. EMBOJ. 11:769-776.

19. Fairman,M.P.,and B.Stillman.1988.Cellular factorsrequired formultiple stages of SV40DNAreplicationin vitro.EMBO J. 7:1211-1218.

20. Goetz, G. S.,F. B.Dean, J.Hurwitz,and S. W. Matson. 1988. Theunwinding ofduplexregionsinDNAbythesimian virus 40 large tumor antigen-associatedDNAhelicaseactivity. J. Biol. Chem. 263:383-392.

21. Hanahan,D.1983. Studiesonthetransformation of Escherichia coli withplasmids. J. Mol. Biol. 166:557-580.

22. Kenny, M. K., S.-H. Lee, and J. Hurwitz. 1989. Multiple functions of human single-stranded-DNA binding protein in simian virus 40DNAreplication:single-strandstabilization and stimulation of DNApolymerasesax and b. Proc. Natl. Acad. Sci. USA 86:9757-9761.

23. Luckow,V.A.,and M. D. Summers.1989.Highlevelexpression of nonfusedforeign genes withAutographacalifornica nuclear polyhedrosisvirusexpressionvectors. Virology170:31-39. 24. Mastrangelo,I.A.,P. V. C.Hough, J.S.Wall,M.Dodson,F. B.

Dean,andJ.Hurwitz.1989.ATP-dependent assemblyof double hexamersofSV40Tantigenatthe viraloriginof DNA replica-tion.Nature(London)338:658-662.

insect cells. J. Virol. 62:3109-3119.

26. Parsons, R., M. E. Anderson, and P. Tegtmeyer. 1990. Three domains in the simian virus 40 core origin orchestrate the binding, melting, and DNA helicase activities of T antigen. J. Virol. 64:509-518.

27. Parsons, R., and P. Tegtmeyer. 1992. Spacing is crucial for coordinationof domain functions within the simian virus 40 core origin of replication. J. Virol. 66:1933-1942.

28. Parsons, R. E., J. E. Stenger, S. Ray, R. Welker, M. E. Anderson, and P. Tegtmeyer. 1991. Cooperative assembly of simian virus 40 T-antigen hexamers on functional halves of the replication origin. J. Virol. 65:2798-2806.

29. Roberts, J. 1989. Simian virus 40 (SV40) large tumor antigen causes stepwise changes in SV40 origin structure during initia-tion ofDNAreplication. Proc. Natl.Acad. Sci. USA 86:3939-3943.

30. Sambrook, J.,E. F.Fritsch,and T.Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

31. SenGupta, D., and J.A.Borowiec. 1992.Strand-specific recog-nition ofa syntheticDNAreplication fork by the SV40 large tumorantigen. Science 256:1656-1661.

32. Tjian,R.1979. Protein-DNA interactionsattheoriginofsimian virus 40DNA replication. Cold Spring Harbor Symp. Quant. Biol. 43:655-661.

33. Wessel, R., J. Schweizer,and H. Stahl. 1992. Simian virus 40 T-antigen DNA helicase is a hexamerwhich forms a binary complexduring bidirectionalunwinding from the viraloriginof DNAreplication. J. Virol. 66:804-815.

34. Wiekowski, M., M. W. Schwarz, and H. Stahl. 1988. Simian virus 40large T antigenDNAhelicase. J. Biol. Chem. 263:436-442.

35. Wobbe, C. R., L.Weissbach, J.A. Borowiec, F. B. Dean,Y. Murakami, P. Bullock, and J. Hurwitz. 1987. Replication of simian virus 40 origin-containing DNA in vitro with purified proteins. Proc.Natl.Acad. Sci. USA 84:1834-1838.

36. Wold,M.S.,andT.Kelly.1988. Purification and characteriza-tion ofreplication protein A, acellularprotein requiredfor in vitroreplicationofsimian virus 40DNA.Proc. Natl. Acad. Sci. USA85:2523-2527.

37. Wold,M.S.,J. J.IA,and T.J.Kelly.1987. Initiationof simian virus 40 DNA replication in vitro: large-tumor-antigen- and origin-dependentunwindingofthetemplate.Proc.Natl. Acad. Sci. USA 84:3643-3647.