THEJOURNALOF BIOLOGICAL CHEMISTRY

Val. 266, No. 8, Issue of March 15, pp. 5062-5071,1991

0 1991 hy The American Society for Biochemistry and Molecular Biology, Inc.

Printed in U.S. A.

Simian Virus40 Large T Antigen UntwistsDNA at the Origin of DNA Replication* (Received for publication, June 20, 1990)

Frank B. Dean and Jerard Hurwitz From the Graduate Program in Molecular Biology, Sloan-Kettering Cancer Center, New York, New York 10021

Simian virus 40 large tumor antigen(SV40 T antigen) untwists DNA at the SV40 replication origin. In the presence of ATP, T antigen shifted the average linking number of an SV40 origin-containing plasmid topoisomer distribution. The loss of up to two helical turns was detected. The reaction required the presence of the 64-base pair core origin of replication containing T antigen DNA binding site 11; binding site I had no effect on the untwisting reaction. The presence of human single-stranded DNA bindingprotein (SSB) slightly reduced the degree of untwisting in the presence of ATP. ATP hydrolysis was not required since untwisting occurred in the presence of nonhydrolyzable analogs of ATP. However, in the presence of a nonhydrolyzable analog of ATP, the requirement for the SV40 origin sequence was lost. The origin requirement for DNA untwisting was also lost in the absence of dithiothreitol. The origin-specific untwisting activity of T antigen is distinct from its DNA helicase activity, since helicase activity does not require the SV40 origin but does require ATP hydrolysis. The lack of a requirement forSSB or ATP hydrolysis and thereduction in the pitch of the DNA helix by just a few turns at the replication origin distinguishes this reaction from the antigen-mediated T DNA unwinding reaction, which results in the formationof a highly underwound DNA molecule. Untwistingoccurredwithoutalag after the start of the reaction, whereasunwound DNA was first detected after a lagof 10 min. It is proposed that the formation of a multimeric T antigen complex containing untwisted DNA at the SV40 origin is a prerequisite for the initiationof DNA unwinding and replication.

Simian virus 40 (SV40)’ provides a model system for the study of DNA replication in mammalian cells. Since it was reported that SV40 DNA replication can be carried out in vitro (1) the proteins shown to be involved now include T antigen, the only viral-encoded protein required (1-3), DNA polymerase a-primase complex (4),topoisomerases I and I1 (5), a eukaryotic single-strand DNA binding protein (HSSB) (6-8), and proliferating cell nuclear antigen (9). Recently, *This study was supported by Grant 5R01 GM34559 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. The abbreviations used are: SV40, simian virus 40; T antigen, SV40-encoded large tumor antigen; Form 11, circular duplex DNA containing at least one single-strand break; Form U, highly unwound circular duplex DNA; SSB,single-strand DNA binding protein; HSSB, human SSB isolated from extracts of HeLa cells; DTT, dithiothreitol; ATPyS, adenosine 5’-O-(thiotriphosphate).

DNA polymerase 6 (10, 11), proliferating cell nuclear antigen, and an additional factor named RF-C or activator 1 (12, 13) have been shown to carry out the elongation of nascent DNA chains, yielding long DNA products from the leading strand template. The addition of a 3’ to 5’ exonuclease, RNase H, and DNA ligase allow the synthesis of full-length covalently closed circular products (14). T antigen binds to thereplication origin and carries out the unwinding of duplex DNA, a step that is required to initiate DNA replication. SV40 T antigen converts a duplex circular DNA containing the SV40 replication origin into a highly unwound form in the presence of SSB, a topoisomerase capable of removing positive supercoils, and ATP (15-17). Unwinding is bidirectional from the origin and occurs on linear as well as circular DNA substrates (18).The reaction occurs efficiently in the presence of Escherichia coli SSB, adenovirus DNA binding protein, herpes virus-infected cell protein 8, and human (HSSB) or yeast SSB (19, 20). Up to 70% of the input DNA can be unwound in the presence of the HSSB (19). Three critical domains in the 64-base pair core SV40 origin have been identified. These sequences, which are essential for replication in vivo and in vitro, have been correlated with the sequences essential for T antigen-mediated unwinding (21, 22). This suggests that the unwinding reaction is a critical aspect of DNA replication. This is also indicated by the finding that the firstspecies of DNA labeled in the SV40 replication pathway in vitro in pulse-chase experiments migrated as an unwound structure (23). Such pulse-labeled DNA can be rapidly chased into replication intermediatesand closed circular duplex DNA products. T antigen from cells infected with a recombinant adenovirus vector is inactive in the unwinding reaction unless treated with a phosphatase such as phosphatase 2& (24), bacterial, or calf intestinal alkaline phosphatase. T antigen isolated from SV40-infected cells or recombinant baculovirus-infected insect cells is active without further treatment (15, 19), although reactions using low levels of these preparations were stimulated 2- to 4-fold by pretreatment of the T antigen with phosphatase.2 Two distinct activities of T antigen combine to allow the unwinding of duplex DNA containing the SV40 origin. One is thehelicase activity (15,25-27); T antigen binds to a region of single-stranded DNA and then translocates in the 3’ to 5‘ direction on that strand,displacing the complementary strand when it traverses through duplex a region. The second activity is an SV40 origin-binding function (28). An oligomeric complex of T antigen assembles specifically at the origin; the assembly is ATP-dependent and the nucleoprotein complex has a bilobed structure (29). The complex protects the complete core origin region from the action of DNase I (30, 31). Scanning transmission electron microscopy revealed that the

5062

F. Dean, P. Bullock, and J. Hurwitz, unpublished results.

5063

DNA Untwisting by TAntigen complex contains up to 12 monomers of T antigen, arranged as a double hexamer and centered symmetrically over the 64base pair core origin sequence (32). These findings suggest that theassembly of a T antigen double hexamer at theorigin is followed bythe initiation of bidirectional unwinding of the duplex DNA. Changes in DNA structure at the SV40 origin induced by T antigen were examined by dimethyl sulfate methylation protection and potassium permanganate (KMn04)oxidation (33). The central origin domain was found to contain numerous T antigen-DNA contacts, whereas the chemical reactivity of the flanking regions indicated that T antigen had melted some of the origin sequences and structurally distorted others. T antigen melts the DNA segment at the early palindrome sequence of the core origin even in theabsence of other origin sequences (34). These studies were recently reviewed (35). The length of a DNA fragment containingthe SV40 origin was not changed by the formation of the T antigen double hexamer (29). This indicated that theDNA was not wrapped around the protein in the complex but that theprotein monomers were positioned around the DNA. This made formation of the complex between T antigen and DNA seem analogous to the binding of restriction endonuclease EcoRI to its DNA recognition sequence, which results in a nucleoprotein complex containing untwisted DNA (36). Since T antigen might initiate the unwinding reaction by untwisting the duplex at the origin, we assayed directly for changes in DNA helical twist induced by T antigen. Changes in DNA helical twist were first measured upon binding of E. coli RNA polymerase by incubating duplex circular DNA containing single-strand breaks, either in the presence or absence of RNA polymerase, and then adding DNA ligase to seal the breaks. After deproteinization, the DNA samples were analyzed by zone sedimentation for a difference in their average topological winding number (37). A refinement and simplification of the technique involved the measurement of the difference in average linking number by gel electrophoresis (38, 39). We show here that T antigen induced a negative shift in the topoisomer distribution of SV40 origin-containing circular DNA, under conditions used to carry out SV40 DNA replication, in the presence of ATP and topoisomerase I. The negative change in topoisomer distribution musthave resulted from a reduction in the helical twist of the DNA in the T antigen-DNA complex at the origin. Compensating positive supercoils were induced in the DNA sequences not bound by T antigen, outside the origin region. Theseunconstrained positive supercoils were removed by the action of HeLa topoisomerase I, lowering the average linking number of the plasmid distribution. Thus, the detection of DNA untwisting by T antigen requires the stable persistance of T antigen complexes containing untwisted DNA. This effect most likely represents a change in the DNA helical twist, because there are two contributors to thelinking number of a duplex circle, twist and writhe (40), and a change in DNA writhe induced by T antigen is unlikely. Any change in writhe that was induced by DNAwrapping around the T antigen complex and was large enough to cause a detectable change in the average linking number would also visibly bend or shorten the length of DNA fragments containing the complex (41). However, no such bending or shortening was observed (29). The SV40 origin-dependent shift intopoisomer distribution by T antigen differs from the DNA unwinding reaction, which yields a highly unwound DNA product, form U, in threeways: (i) only a small number of helical turns areremoved, (ii) ATP hydrolysis is not required, and (iii) an SSB is not essential.

Therefore we refer to this reaction specifically as DNA untwisting to distinguish it from the DNA unwinding reaction. We suggest that the formation of the T antigen complex containing untwisted DNA at the origin is a prerequisite for the initiation of DNA unwinding by T antigen. The shift in topoisomer distribution and the sensitivity to potassium permanganate oxidation (33) are two complementary means for analyzing DNA untwisting by T antigen at the replication origin. Potassium permanganatesensitivity reveals which particular DNA sequences become structurally altered, whereas the change in topoisomer distribution shows the magnitude of the untwisting. The two assays have been used to examine the influence of factors, such as time, ATP concentration, and temperature onDNA untwisting by T antigen,3 and show that theT antigen-DNA interactionat theorigin goes through a multi-step process. We have noted previously that T antigen untwists SV40 origin-containing DNA (32, 43). Earlier investigations of the untwisting of duplex DNA by T antigen were inconclusive (44, 45). Roberts has reported that T antigen changed the topology of a small circular duplex DNA containing the SV40 origin in a reaction that required ATP hydrolysis (46). He concluded that T antigen formed three specificcomplexes with the SV40 origin which contained DNA untwisted by -1, -2, and -5 turns. However, the studies presented here suggest that both small and large circular DNA molecules containing the SV40 origin are untwisted by an average of -1.5 turns and this is not in accord with the results reported by Roberts (46). MATERIALS ANDMETHODS

Enzymes and DNA-T antigen was purified from SV40 cs1085infected cos cells or recombinant baculovirus vEV55SVT-infected Sf9 cells (47) as described previously (32). Preparations of HeLa topoisomerase I (15) and HSSB (19) were as described, E. coli topoisomerase I and DNA gyrase were gifts from Dr. Kenneth Mariansof this Institute. DNA substrates, pSVOlAEP (2792 base pairs in size, referred to as ori' in this paper), pBR322AEP (referred to as ori- in this paper), and pSVLD, 10 kilobase pairs in size ( E ) , have been described. Plasmids pON-WT (48), OR4, OR1 (49), Del 5238-6, Bal 5218, and Bal 23 (22) were gifts from Dr. Peter Tegtmeyer (State University of New York, Stony Brook). Relaxed plasmid DNA, used as a substrate for untwisting reactions, was prepared as described (15). The 32P-labeled678-base pair circular DNA was constructed by first incubating3.6 pg ofori+ DNA as a substratefor DNA replication. The reaction mixture (0.6 ml) containing 40 mM creatine phosphate (di-Tris salt, pH 7.7), 7 mMMgC12, 1mM DTT, 4 mM ATP, 14 pg of creatine kinase, 200 pM each of CTP, GTP, and UTP, 100 p~ each C T P cpm/pmol), of dATP, dGTP, and dTTP,25 p~ [ C X - ~ ~ P ] ~(13,000 10 pg SV40 T antigen, and cytoplasmic extract (1.6 mg of protein) prepared from HeLa cells was incubated at 37 "C for 2.5 h (3). The radiolabeled ori+ DNA reaction products (3 X lo6 cpm/pg) were cleaved by restriction endonucleases NdeI and Tag1 (New England Biolabs), and thefour resulting fragments were separated by electrophoresis through a 5% polyacrylamide gel. After autoradiography, the region containing the 678-base pair origin-containing fragment was excised from the gel. The gel slice was crushed and theDNA extracted with a sterile solution of 0.5 M ammonium acetate, 1 mM EDTA, and 0.1% sodium dodecyl sulfate. The DNAwas collected by ethanol precipitation and circularized by incubation at 0.4 pg/ml in a reaction mixture (1.0 ml) containing 50 mM Tris.HC1, pH 7.5, 10 mM MgC12, 1 mM ATP, 5 mM DTT, and 2 units of T4 DNA ligase at 15 "C for 12 h. The yield of monomer covalently closed circular DNA wasapproximately 75%. The DNAwas concentrated under vacuum using a Speed Vac (Savant) andethanol-precipitated. One-half of the circular DNA product (0.2 pg) was supercoiled in a reaction mixture (0.1 ml) containing 50 mM Tris.HC1, pH 7.5, 18 mM potassium phosphate, ~

~~

Borowiec, J. A., Dean, F. B., and Humitz, J. (1991) J. Virol., in press.

DNA Untwisting by TAntigen

5064

A. p H 7.5, 10 mM MgC12.5 mM spermidine, 5 mM DTT, 1 mM ATP, 50 Tmtvn - + - + + pglrnl of bovine serum albumin, and 2 units of DNA gyrase (50). A T P - - + + Incubation was for 30 rnin a t 30 "C and the DNA was collected by ethanol precipitation. DNA UntwkfingAssay-Standard reaction mixtures (30 pl) containing 40 mM creatine phosphate (di-Tris salt, pH 7.7), 7 mM MgCI,. 1 mM DTT, 4 mM ATP, 10 units of HeLa topoisomerase I, 0.9 pg of -A bovine serum albumin, 0.3pg of relaxed ori' DNA, and the indicated amounts of T antigen were incubated for 45 min a t 37 "C. Reactions 0 were terminated by the addition of a solution (5 pl) containing 2% - T 0°C-n , . sodium dodecyl sulfate, 0.1 M EDTA, 1 pg of glycogen, and 5 pg of proteinase K.Due to the temperature dependence of the helix rotation '-. -" angle of duplexDNA(Sl),care was takentoavoidtemperature +ToM-gen , ' ' changes. After incubation a t 37 "C, the solution used to terminate the reaction was added rapidly and the reaction mixtures were immedi. '- L ately incubated a t 37 "C for 30 min more. Samples were precipitated FIG. 1. Requirements for DNA untwisting by T antigen. A. with ethanol, resuspended, and analyzed by electrophoresis through reactions containing 0.75 pg of T antigen and ATP, as indicated. agarose gels(1.8%) containing chloroquine phosphate (2-4 pg/ml) were as described under "Materials and Xlethods." Gel electrophoresis (52)inboththe gel andtherunningbuffer(Tris-acetate-EDTA, TAE). Samples analyzed by electrophoresis in gels containing chlo- was in thepresence of 2 pg/ml chloroquinephosphate. Form I I . nicked circular DNA; Relaxed, covalently closed circular topoisorners: roquinemust be free of sodium dodecyl sulfate in order to avoid anomalous migrationof the circular DNA. After electrophoresis, gels A, change in average linking numher with respect to l a n ~1. 19. densitometric scans of lanes 1 (upper trace) and 4 (lower trace). werewashedfor1h in 50 mM NaCl to remove thechloroquine, stained for 1 h in TAE buffer containing0.5 pg/ml ethidium bromide, Ckl and photographed under UV transillumination. Photographic nega5238 eo1 B o 1 DNA On+ OR4 OR1 - 6 &I5218 23 NN-WT tives were scanned using an LKB 2202 Ultrascan laser densitometer. r . l The amount of DNA in the bands of an individual lane was deterTontlgen - + - + - + - + - + -+-+ -+ mined by integration using the LKB 2190 Gel scan program on an Apple IIe computer equipped with an arithmetic processor. Each band was assigned an integral value to represent the linking number of that topoisomer. The fraction of the total DNA in each band was multiplied by its integral value and the products were summed togive the (relative) average linking number of the DNA in that sample. The difference intheaveragelinkingnumber (A) between individual experimentscould then he calculated. -A 1: ,; 1'4 335 '..,, - 2 6 DNA Unwinding Assay-Unwinding reactions were carried out by FIG. 2. DNA sequence requirements for DNA untwisting by the procedure described above except that reactions contained 0.36 T antigen. Reactions contained0.75 pg of T antigen where indicated. pg of ori' DNA, 0.4 pg of T antigen, and 1.0 pg of HSSB toallow the Del 52.18-6 is a deletion of nucleotides .S238-6 encompassing the two production of the unwound Form U DNA. Samples were analyzed by electrophoresis through agarose gels (1.5%) containing 1 pg/ml chlo- central pentanucleotides. Bal5218 and Ral2.3 are deletions (produced by Bal 31 exonuclease digestion) into the early side of the origin, roquine and quantitated by scanning densitometry (19). of the origin. stopping DNA Helicase Assay-Reaction mixtures contained40 mM creatine stopping a t position 5218, and into the late side phosphate (di-Tris salt, pH 7.7), 7 mM MgC12, 1 mM DTT, 0.2 pg T a t position 23, respectively (22). pON-WT contains a 19-base pair insert homologous to nucleotides5191-5209 of the SL'40 sequence ( T antigen,ATPasindicated,and15 fmol of DNAsubstrate.The antigen binding site I). substrate, a "P-3' end-labeled32-nucleotide-long oligonucleotide (nucleotides 5156-5125) hybridized to single-stranded circular 6x1'74 DNA, was a gift of Mr. Y. S. Seo of this laboratory. Incubation was ent topoisomer distributions was calculated (see "Materials for 30 min a t 37 "C; reactions were terminatedandanalyzedas and Methods").Fig. 1R showsa comparison of the laser scans described (27). _ L

._

~

I-

-".-

"

""-

':

of the 1st and 4thlunes in Fig. lA. Thus, T antigen removed about 1.3 helical turns from the DNA. DNA Sequence and Nucleotide Requirements for the UnT AntigenUntwists Plasmid DNA Containing the SV40 twisting Reaction-The SV40 core origin alone was sufficient Origin-To detectDNAuntwisting,relaxedSV40originto allow untwisting (Fig. 2). Plasmid pOR1, containing only containing (ori') DNA was incubated a t 37 "C with T antigen, the SV40 sequences comprising the 64-base pair core origin HeLa topoisomerase I, and ATP. After deproteinization and of replication, was as effective as pOR4 andori' DNA, which gel electrophoresis, the distributionof relaxed plasmid topo- containthecoreas well astheflanking SV40sequences, isomers was shifted (Fig. lA, lane 4). There were no changes including T antigen binding site I. Plasmid Del 5238-6, conin the distribution of topoisomers in the presence of T antigen taining a deletion of the two central pentanucleotide T antigen or ATP alone. TheDNAwasrelaxedunderthereaction binding repeats from the SV40 origin, was ineffective, as was conditions, and the presence of chloroquine in the electropho- ori- DNA. Plasmids Bal5218 and Ral23, containing deletions a few pos- intotheearlyandlatesides resis buffer caused the topoisomers to accumulate of thecore originsequence, I1 plasmid respectively, showed a 4-fold reduction in the degree to which itive supercoils and to migrate faster than the form DNA during electrophoresis (52). Since the slower-migrating they were untwisted. Plasmid pON-WT, containing T antigen topoisomers have a lower linking number than the fasterorigin sequences, was not bindingsite I butlackingany migrating topoisomers,T antigen reduced the average linking untwisted. Thus, the central T antigen binding sequences of number of the topoisomer population. Topoisomerase I was the core origin arerequiredfor DNA untwisting, whereas required to observe the effect (data not shown), indicating disruption of the flanking core origin sequences results in a as a topoisomerase. 4-fold reduction in untwisting. In comparison, disruption of that T antigenwasnotfunctioning Rather, this showed that T antigen was directly untwisting the central T antigen binding sequencesor either of the two the DNA helical structure. flanking core origin domains completelv abolished the SSRTo quantitate the extentof untwisting, photographic neg- dependentDNAunwindingactivitv(22). DNA sequences atives were scanned usinga laser densitometer and the differ-outside of the core origin had little effect on the untwisting ence in the average linking number (A) between differ- reaction, as was also true for the DNA unwinding reaction RESULTS

DNA Untwisting by T Antigen carried out using purified proteins (22). The effectiveness of various nucleotides in supporting the DNA untwisting reaction was tested (Fig. 3A). The most effective, based on the nucleotide concentration that yielded half-maximal untwisting, was dATP, followed byATP, dTTP, and UTP. CTP and GTP more than 40-fold less effective than ATP. The nucleotide cofactor specificity of the untwisting reaction is the same as thatof the DNA helicase and the duplex unwinding reactions of T antigen (27), but different from that of its intrinsic RNA helicase activity (53). The dependence of DNA unwinding (the formation of a highly unwound duplex DNA circle) and DNA helicase (the displacement of an oligonucleotideannealed to a single-strand circular DNA) activities of T antigen on the concentration of ATP were also examined. The concentration of ATP required for efficient DNA unwinding was more than 10-fold higher than the concentrations required for the DNA helicase or DNA untwisting reactions (Fig. 3B). When present at a concentration of 4 mM, both dATP anddTTP substituted for ATP in supporting DNA unwinding. However, they were also ineffective at concentrations of 1 mM or lower (datanot shown). The unwinding of duplex DNA by T antigen is stimulated by phosphate ions such as creatine phosphate or inorganic phosphate; this dependence is not due to a requirement for the regeneration of ATP (27). However, the omission of creatine phosphate had no effect on the DNA untwisting reaction (data not shown). DNA untwisting was sensitive to salt; 50 mM NaCl reduced the extent of untwisting by 50%, and 100 mM NaCl inhibited untwisting completely (data not shown). Influence of T Antigen Concentration on Untwisting-In the presence of ATP, the extent of untwisting increased as the concentration of T antigen was raised (Fig. 4). Only a low level of untwisting not dependent on the SV40 origin was observed. In the presence of 2.0 pg of T antigen, the distribution of ori+ topoisomers was shifted -1.9 turns, of which

5065 A.

Ori+DNA

Tantigen(pg)

0 0.2 0.4 0.6 0.9 1.21.6 2.0

B. Ori-DNA T antigen( p g ) 0 0.2 0.4 0.6 0.9 1.21.6

c.

2 x

A

’

2.0

!: 0.5

0.4

0.8

1.2

1.6

2.0

T antigen (pg)

FIG. 4. Influence of T antigen concentration on untwisting. Agarose gel electrophoresis of reactions containing ( A ) ori’ DNA or ( B ) ori- DNA and T antigen as indicated. C, quantitation of the results using ori’ (0)or ori- DNA (0).

-1.6 turns were origin-specific. However, the distribution of topoisomers was broader in the presence of larger amounts of T antigen. This indicated that theuntwisted DNA consisted of a population of molecules that were untwisted to varying A. extents. 1.2 No evidence of a plateau in the untwisting reaction was observed, even though T antigen was present in great excess. 0.9 In a reaction containing 1.2 pg of T antigen the ratio of T A antigen monomers to plasmid DNA molecules (containing a 0.6 single SV40 origin) was 1OO:l. a I One reason for the failure of the extent of untwisting to 0.3 reach a plateau might have been that a fraction of the T n antigen was inactive due to phosphorylation at serine residues 10 100 1000 10000 (24). In fact,T antigen isolated from recombinant adenovirus[Nucleotide] (pM) infected cells was inactive in the untwisting reaction, as it is B. in DNA ~nwinding.~ However, the extent of untwisting also continued to increase as the concentration of baculovirus T antigen pretreated with phosphatase 2& was increased (data not shown). Another possible explanation for the failure of the untwisting reaction to level off was that the helicase activity of T antigen increased the signal by causing limited unwinding of the duplex outward from the origin. To avoid this possibility OQ . . . _ . _ ’ ~ ~ ~ ’ . ’ ’ ~ -10 ’ a titrationof T antigen was carried out using ATP-yS in place 10000 1000 10 100 of ATP, but theresults were surprising (Fig. 5). The increase [ATPI W ) FIG.3. Nucleotide requirements for DNA untwisting and in the extent of untwisting observed was linear with respect DNA unwinding by T antigen. A, untwisting reactions contained to the amount of T antigen added, but the requirement for 0.75 fig of T antigen, and the concentration of various nucleotides the SV40 origin was lost. The untwisting of a plasmid conwere varied as indicated. Untwisting was quantitated as described taining the origin was only 2-fold greater than untwisting under “Materials and Methods.” B, DNA helicase activity (0)and observed using a plasmid lacking the origin. The same result v

-

U

’

DNA unwinding activity (0)were assayed as described under “Materials and Methods,” with the concentration of ATP varied as indicated.

F. B.Dean and J. Hurwitz, unpublished results.

DNA Untwisting by T Antigen

5066 A.

'

A.

or) -

ori+

DNA

Tarlil,,gl

Tantigen rlmc (rnln)

0 0.306 09 1.2 1.6 12 12"O 0 9 1.6' 04 0 8

Form Il

"

0

1

"

2

5

+ + + + + + + 1

2

5

1554456060

- I

If '0

T anwen

tions containing ori' DNA, T antigen, and HSSB as indicated. ATP was omitted from the reactions and A T P r S (4 mM) was added. R, quantitation of the results inA using ori- DNA (0)or ori- DNA (0).

of T antigen, as indicated, were assembled on ice and transferred for each to a 37 'C bath to initiate the reaction. Incubation time reaction is shown. R , quantitation of the results shown in A .

t

0

0

60

pg

" "

30 45 Tme (mm)

FIG. 7. Time courseof untwisting. A. reactions containing0.72

Ora

FIG. 5. Influence of T antigen concentration on untwisting in the presenceof ATPrS. A , agarose gel electrophoresis of reac-

lwr

15

0.3

0.6

0.9

1.2

1.5

1.8

DNA ( m )

FIG. 6. Influence of DNA concentration on DNA untwisting. Reactions containing 0.72 pg of T antigen were as described under "Materials and Methods," except that the amount of ori' DNA was varied as indicated. Two identical reactions containing 0.15 pg of DNA were carried out and then combined. After precipitation of the products with ethanol, DNA was resuspended and an identical amount of DNAfrom eachsample (0.3 p g ) wasanalyzed by gel electrophoresis to quantitate untwisting. A in the presence of 0.15 pg of DNA was defined as 100% and was equal to-0.8 turn.

pg

10

20 Time (min)

30

FIG. 8. Stability of DNA untwisting. Reactions containing0.75 of T antigen were incubated for 30 min a t 37 'C, after which time,

4.8 pg pSVDR31 plasmid DNA was added as a competitor, and the incubation continued a t 37 'C for the times indicated10).Competitor DNA was omitted from parallel reactions (0).

ature shift altered theDNA twist, and the changein topoisomer distribution over 5 min in the absence of T antigen (Fig. 7A, lanes 1-4) showed that topoisomerase I broughtthe plasmid population completely to itsnew equilibrium position was observedwhen the reaction mixtures contained 1 mM within 2 min. A rapid increase in untwisting carried outby T ATPyS and2 mM DTT. Thus, the bindingof an ATPanalog antigen was observed up to15 min. at which time thereaction of untwisting observed that is not hydrolyzed by T antigen5 allows T antigen to carry was almost complete. Thus, the extent out non-sequence-specific untwisting, in addition to origin- after 30 or 45 min of incubation was the full yield possible specific untwisting, whereas ATP hydrolysis limits the effect under those conditions. No significant lag in untwisting was of a lag before specifically to the SV40originsequence. The presence of detected. This is in contrast to the observation HSSB had no effect on untwisting mediatedby ATPyS (Fig. initiation of the unwinding reaction, asdescribed below. Effect of Added Competitor DNA-The stability of the T 5A, compare lane 5 with lanes 7 and 8). Influence of DNA Concentration on Untwisting-In the antigen-DNA complexes responsible for the untwistingeffect presence of a fixed amount of T antigen (0.72 pg), the extent was determined. Untwisting reactions were incubated for 30 of DNA untwisting was reduced as the DNA concentration min at 37 "C, after which 4.8 pg of competitor DNA was was increased and the molar ratioof T antigen monomer to added, along with additionaltopoisomerase I, and incubation DNA decreased from 360 to 30 (Fig. 6). This suggests that was continued at 37 "C for the times indicated(Fig. 8). Comat petitor DNA was omitted from parallel reactions. The comassembly of the T antigen complex that untwists the DNA the origin isnot highly cooperative and also that the T antigen petitor DNA, pSVDR31, is an 8-kilobase pair plasmid conmolecules bound at theorigin may be in equilibrium with the taining two SV40 origin sequences/molecule, so that the 4.8 pg of competitor DNA contained a 10-fold molar excess of unbound molecules in solution. Time Course of Untwisting-The effect of incubation time origin sequences over those present in the 0.3 pg of pSVOl on untwisting was investigated. Reactions were initiated by EP substrate. The competitor migrated more slowly through transfer from 0 to a 37 "C water bath and were terminated the agarose gel than the 2.8-kilobase substrate DNA and did of untwisting.In the after incubation for the times indicated(Fig. 7). The temper- notinterferewiththequantitation absence of competitor the extentof untwisting was stable, as b J. Hurwitz, unpublished results. expected. However, the addition of competitor caused a de-

DNA Untwisting by T Antigen A +HSSB Tantigen Time(mln) 5 5 10 15 20 3045 60

'

+++++++

B

-HSSB

-+++++++

A

30 I

On+ DNA

Tmlgm

DTTImM)

5 5 lO15m304560

B

5067

-++++++0 0 0205 1 2

4

2

0 r 1 - DNA

Tmli - + + + + + + DTTImM) 0 0 0 2 0 5 1 2 4 2

lime (min)

C. Tanti DTTfmMI

Vnwtnd~nq

-++++++ 0

00205

1

2 4

0.4

0.2

"5

15

30 Time (min)

45

60

FIG.9. Comparison of DNA untwisting and DNA unwinding. A, unwinding reactions containing 0.4 pg of T antigen were incubated for the times indicated in the presence of 1.0 pg of HSSB and electrophoresed in the presence of chloroquine (1.0 pg/ml) to allow analysis of both unwinding anduntwisting. Form U, highly unwound covalently closed circular DNA. R, untwisting reactions were identical to the reactions in A except for the omission of HSSB. C, quantitation of the unwound DNA, Form U, formed in the reactions shown in A. D, quantitation ofuntwisting generated in the presence of HSSB (0)or in the absence of HSSB (0).

crease in untwisting. Untwisting dropped by about 50% after 15 min in the presence of competitor. This indicated that the T antigen-DNA complexes that cause the DNA untwisting are not stable. Addition of the competitor along with the substrate at the startof the reaction reduced the untwisting observed by 90%. Comparison of DNA Untwisting and DNA Unwinding by T Antigen-To clarify the relationship between untwisting and unwinding in SV40 DNA replication, a time course was performed in the presence of HSSB and both activities were measured (Fig. 9A). In addition, the time course was carried out in the absence of SSB (Fig. 9B). Quantitation of the unwinding reaction revealed a 10-min lag before DNA unwinding was observed (Fig. 9C), whereas the untwisting reaction showed no such lag, either in the presence or absence of SSB (Fig. 9D). This suggests that the T antigen-DNA complex containing untwisted DNA is established quickly but that a slow step then follows before the DNA unwinding reaction is initiated. However, once initiated, unwinding progresses rapidly, at a rate greater than one kilobase per min.6 In the presence of SSB thedegree of untwisting was reduced approximately 20%. Influence of DTT Concentration on DNA Untwisting and Unwinding-During the course of these experiments it was noted that DNA untwisting by T antigen was enhanced at low concentrations of DTT, whereas DTT concentrationsof 1 mM or more permitted a reduced but consistent signal (Fig. 1OA). Furthermore, in the absence of added DTT, untwisting lostits SV40 origin specificity, and ori- DNA became a M. Dodson. F. Dean, and J. Hurwitz, unpublished results.

FIG. 10. Influence of DTT concentration on DNA untwisting and unwinding. Untwisting reactions containing 1.2 pg of T antigen and ori' DNA ( A ) or ori- DNA ( R ) werecarriednut as described, except that the concentration of DTT was varied as indicated, and gel electrophoresis was carried out in the presence of 4 pg/ ml chloroquine phosphate. C, unwinding reactions were carried nut as described except the DTT concentration was varied as indicated. D, quantitation of untwisting as a function of DTT concentration using ori' (0)or ori- DNA (0).E , quantitation of DNA unwinding as a function of D m concentration.

substrate (Fig. 10B). The ommission of DTT from the reactions reduced its concentrationto below 0.1 mM; the residual DTT was contributed by the DTT present in the T antigen and topoisomerase I preparations. DNA unwinding required DTT (Fig. lOC), and quantitation of these results showed that thelevels of DTT required to abolish origin-independent untwisting were identical with the levels required for DNA unwinding (Fig. 10, D and E ) . However, origin-specific untwisting probably does not require DTT, asseen by the greater untwisting of ori+ DNA even in the absence of DTT. The importance of sulfhydryl residues in the activity of T antigen was also indicated by the sensitivity of ATPase, helicase, DNA unwinding, and DNA replication assaysto pretreatment of T antigen with N-ethylmaleimide, although the binding of T antigen toori' DNA was insensitive (15). The untwisting reaction in the absence of DTT was approximately linear with respect to T antigen concentration, both with ori' and ori- DNA (Fig. 11).The reaction did not reach a plateau and the removal of up to five helical turns was observed using ori+ DNA, consistent with the loss of origin dependence in the reaction shown inFig.10. In the absence of DTT, the untwisting of ori- DNA ranged from 25% of the untwisting observed with ori' DNA (Figs. 10 and l l ) , to as high as 50% of the untwisting observed with ori' DNA (Fig. 12). This is similar to what was observed in the presence of ATPrS (Fig. 5). In contrast, in the presence of DTT the untwistingof ori+ DNA was a t least 20-fold higher than theuntwisting of ori- DNA, when compared in the linear range with respectto the amount of T antigen in the reaction (Figs. 2 and 4). In the presence of DTT, at high levels of T antigen, the untwisting of ori' DNA was 5-fold more effective

DNA Untwisting by T Antigen

5068

T antigen

(re)

FIG. 11. Influence of T antigen concentration on untwisting in the absenceof DTT. Reactions containingT antigen as indicated and usingori' DNA (0)or ori- DNA (0)were carried outas described except that DTTwas omitted. After electrophoresis through chloroquine-containing (4 pg/ml) agarosegels, untwisting was quantitated as described. A.

on+

DNA

0.

ora-

-++++-++++

Tonlqon NoCllrnM)

0

0 2 0 4 0 8 0 0 0204080

51

NsCl (mM1

FIG.12. Influence of NaCl on untwisting in the absence of DTT. A, reactions containing1.2 pg of T antigen were carried out as described, withtheaddition of NaCl as indicated,andproducts electrophoresed through an agarose gel containing 4 pg/ml chloroquine phosphate. E , quantitation of the results using ori' DNA (0) or ori- DNA (0).

electrophoresis. There was also one minor topoisomer that migrated slightly slower than form I1 (Fig. 13. lane 4 ) . The effect of temperature on DNA twist (51) and the temperature difference between the ligation (15 'C)and gel electrophoresis (25 "C) conditions causes the introduction of one positive supercoil during gel electrophoresis. In addition, these topoisomers were unaffected by treatment with E. coli topoisomerase I (Fig. 13, lanes 1-3). which is not active on positively supercoiled DNA. This indicates that the major topoisomer had one positive supercoil under the electrophoresis conditions and the minor one had no supercoils. Upon treatment with HeLa topoisomerase I, the topoisomer population shifted so that most of the DNA had no supercoils, and asmall proportion migrated as a topoisomer that we suggest had -1 supercoil (lane 5). Upon incubationin the presence of T antigen and ATP, the distribution shifted so that themajority of the DNA migrated as topoisomers that possessed -1, -2, and -3 supercoils (lane 6 ) . This indicated that T antigen shifted the topoisomer distribution of the 678-base pair circle by about -1.5 turns,consistent with the resultsobtained using the 2800 base-pair pSVOlAEP. The addition of HSSB resulted in theformation of an additional faint band migrating slightly faster through thegel than -3 (lane7 ) . To check the assignment of the number of supercoils, the substrate DNA was treated with DNA gyrase, which resulted in three topoisomers running in positions -1, -2, and -3 (lane 9 ) . Treatment of the supercoiled DNA with E. coli topoisomerase I demonstrated that theDNA could be relaxed to the -1 and 0 positions, showing that no other intermediate topoisomers were produced by topoisomerase I upon relaxation of the topoisomer designated -3 (Fig. 13, lanes 10-13) and providing a positive control for the activity of the E. coli topoisomerase I. This confirmed the topoisomer assignments and proved that no species were missing from this accounting. The supercoiled DNA was also boiled, which removed the bands of Form I1 and linear DNA, and resulted in two new bands (lane 8).The faster-migrating band co-migrated with the novel band produced in the presence of HeLa SSB (lane 7 ) .This is consistent with that band being a highly unwound and irreversibly denatured circular form. Ligalm Rodua

Substrok DNA

than the untwisting of ori- DNA (Fig. 4). Effect of Salt on Untwisting-We examined whether NaCl would preferentially reduce origin-independent untwisting in the absence of DTT. Increasing amounts of NaCl inhibited DNA untwisting in theabsence of DTT (Fig. EA). However, NaCl inhibited the untwisting observed using both ori+ and ori- DNA to a similar extent, and concentrations of greater than 80 mM NaCl inhibited both reactions completely (Fig. 12B; data not shown). In addition, NaCl inhibited the DNA unwinding reaction (data notshown). In theabsence of DTT, T antigen binding to nonspecific as well as origin-specific DNA sequences showed the samesensitivity to NaCI. Untwisting of a 678-Base Pair Duplex DNA Circle-Recently it was reported that T antigen untwisted a 572-base pair circular DNA with the formation of three distinct nucleoprotein complexes, containing DNA untwisted by -1, -2, and -5 turns, respectively (46). In contrast, we had no evidence for three complexes of T antigen using a2800-base pair substrate. Todirectly assessthe effect of the size of the SV40 origin-containing substrate on the DNA untwisting reaction, we tested the effect of T antigen on a 678-base pair circle, similar in size to the oneused previously (46). The 678-base pair circular DNA substrate existed as one major topoisomer that migrated faster than form I1 upon

Boi HeLa Tow I TOllli

I

+++ -++ +

HSSB

E.&~TopoI(ngl

supercoiled

+

1

"

8

4

0 0.5 1 2 4

2 0

1 2 3 4 5 6 7 8 9 l o 1 1 1213 . , .-

-

O-

-1 +1-2

m

m

o

m

*

~

f

i

;

-FwrnD ~ ~

~

~

~

4

-3-

liaar-

'

"0""

"".

FIG. 13. Untwisting of a 678-base pair circular origin-containing DNA by T antigen. Substrate preparationwas as described under "Materials and Methods." Reactions containing E. coli topoisomerase I, as indicated (lanes 1-4 and 9-13), were incuhated in the presence of 10 mM Tris.HCI, pH 7.6, 10 mM KC1, 6 mM MgCI?, 0.5 mM DTT, and 50 pg/ml bovine Serum albumin for 30 min a t 37 'C. Untwisting reactions (lanes 5-7) contained 1.2 pg of T antigen and 1.0 pg of HSSB, as indicated. Reactions contained 2 ng of labeled 678-base pair substrate DNA and 0.3 pg unlabeled on+ DNA so that used in the the total T antigen-DNA ratio was comperahle with that other reactions. Products were analyzed by electrophoresis through 3.5% polyacrylamide gels (46). An autoradiograph of the dried gel is shown.

5069

D N A Untwisting by T Antigen DISCUSSION

In the presence of various nucleotides, T antigen removed up to two helical turns of the DNA helix at the SV40 core origin sequence. These results differ with the report that T antigen forms a complex that specifically untwists a small circular DNA at theorigin by -5 turns (46). We have verified and -3 to the the relative assignments of +1, 0, -1,-2, topoisomer bands that were formed in our experiments with a small circular ori+ DNA (Fig. 13). Withoutsuch verification, we have reservations about the assignment of -5 to thefastest migrating topoisomer band by Roberts (46). Our results would be in accord with those of Roberts if the band indicated to have -5 turns had -3 instead. In addition, the suggestion that T antigen makes three distinct complexes with the sV40 origin (46) resulted from the interpretation that each new topoisomer produced after a shift inthe distribution resulted from a distinct interaction between T antigen and the DNA. We believe that the appearance of individual topoisomers cannot be interpreted this way, but that a valid analysis involves quantitation of the shift in the topoisomer distribution. Our experiments show that untwisting can be measured using a small circle as well as with a 2.8-kilobase pair plasmid DNA. However, a small circle has a fewer number of bands in its topoisomer distribution, so it is more difficult to calculate thecenter of the distribution, andtherefore the extent to which the distribution is shifted under various conditions. Thus, the use of a small circle will not be advantageous for future studies of the effectiveness of mutated T antigen or SV40 origin sequences in the DNA untwisting reaction, as suggested by Roberts (46). T antigen also untwisted pSVLD, a 10-kilobase pair plasmid containing a single SV40 origin, the extent as the other substrates by -1.5 turns, or about same (data not shown). We conclude that the size of the origincontaining plasmid DNA, in the range of 0.6-10 kilobases, has no effect on the extent towhich T antigen untwists the DNA at the SV40 origin. The removal of up to two helical turns by T antigen at the origin is consistent with the size of the regions of DNA structural distortion induced by T antigen as determined by potassium permanganate oxidation (33,42). Those studies showed that a DNA region spanning 8 base pairs on the early side domain of the SV40 origin was melted, and theDNA was structurally distorted, although not melted, over a region of 16 base pairs on the late side domain of the origin, in the A/ T-rich region. The central domain of the origin, containing the four pentanucleotide T antigen binding sites, was not structurally distorted. It is likely that the structural changes in thetwo flanking domains of the origin each result from the removal of approximately one helical turn from the DNA. The changes in each region are likely due to theaction of one of the lobes of the bilobed T antigen double hexamer (32,35). Untwisting and the T Antigen Double Hexamer-The T antigen-mediated DNA untwisting reaction and theformation of the previously identified bilobed double hexamer (29, 32) are probably related. Both the untwisting reaction and double hexamer formation require ATP and origin-containing DNA. Also, the untwisting reaction requires the SV40 core origin sequence and the double hexamer covers the core origin (29, 32). Neither complete double hexamer formation nor untwisting occur at 0 "C; both reactions require higher temperature (32,42). T antigen canassemble into hexamers in the absence of DNA and in the presence of ATP; this occurs without a detectable timelag and is not affected by the DTTconcentra-

T anligen

Mxomers

+

Free DNA

T anhgen "

11-mr

-

Unwinding T antigen

1

12-mer

4

2

Initiates

FIG. 14. Scheme for initiation of unwinding by SV40 antigen. See "Discussion" for details.

T

tion (32).7 Thus, double hexamer formation at the origin probably also initiates without a noticeable lag, as does the untwisting reaction. The DNA untwisting reaction is likely a prerequisite for the DNA unwinding reaction. Initiation of helicase activity requires the binding of T antigen to single-stranded DNA. The double hexamer likely disrupts the duplex to generate stable single-stranded regions. This is manifested both by DNA untwisting andby the sensitivity of the nucleotide bases to chemical modification (33). Unwinding might then be initiated as the two hexamers at the origin move apart and initiate DNA strand displacement activity. Alternatively, T antigen molecules free in solution could bind to the singlestranded regions to initiate unwinding. The Role of Untwisting in Initiating the Unwinding Reaction-The untwisting reaction differs from the unwinding reaction in several properties. First, there is no lag before untwisted molecules accumulate, whereas there is a 10-min lag beforeunwinding begins. Second, there is no DTT requirement for untwisting, whereas there is for unwinding. Third, there is no stimulation by a phosphate compound such as creatine phosphate for the untwisting reaction, as thereis for unwinding. Fourth,untwisting occurs efficiently at 25 "C, whereas unwinding requires higher temperature (15, 42). Finally, DNA unwinding requires a 10-foldhigher concentration of ATP than the untwisting reaction does. Thus, several conditions which favor untwisting do not support unwinding. We suggest that theassembly of the T antigen-DNA complex containing untwisted DNA is not the rate-limiting step for the initiation of unwinding. Rather, the rate-limiting step is subsequent to untwisting. This suggestion is illustrated by the scheme shown in Fig. 14. T antigen monomers free in solution assemble on the DNA sequences at the sV40 origin in a number of steps. Partial nucleoprotein structurescan probably untwist the DNA to some degree. The addition of the last monomer of T antigen to theprotein-DNA complex (step 1) could be a slow step in the scheme, in which case the double hexamer would rapidly initiate unwinding. Alternatively, the step after formation of the double hexamer (step 2) could be the slow step, implying arequirement for a structural or confirmational change in the fully assembled double hexamer before unwinding could initiate. The second alternative has a precedent in a model proposed for the mechanism of action of the restriction endonuclease EcoRI (54). It was suggested that after formation of a specific complex between EcoRI and its DNA recognition sequence, a conformational change occurs in the protein which activates the cleavage center of the enzyme. Parallels with the recA-DNANucleoprotein StructureThe structural alteration induced by T antigen within the SV40 origin sequence is substantial. The removal of two turns represents a 720" rotation of one end of the 60-base pair sequence relative to the otherend. In contrast, the binding of restriction endonuclease EcoRI to itsrecognition site untwists the DNAby25" (57), whereas DNA binding by the lac repressor untwists the helix by 55" (58).Thus, it is unlikely that T antigen and EcoRI untwist DNA in a similar manner. Perhaps a better analogy for the structure assembled by T antigen is the nucleoprotein complex formed between the E. F. B. Dean and J. Hurwitz, manuscript in preparation.

5070

DNA Untwisting by T Antigen

coli recA protein and duplex DNA in thepresence of ATPyS (59). The rec A protein assembles a helical structure around single-stranded or duplex DNA. Complex formation is nucleotide-dependent, and the structure formed in the presence of ATPyS is more stable than theone formed in the presence of ATP (61). Two turns of the rec A helix contain about 12 rec A monomers and cover 195 A (62). This would coverabout 57 base pairs of B-form DNA, a region similar to the size of the SV40 core origin. The DNA in the rec A nucleoprotein structure is untwisted on average by 15" per base pair (63), which would be equivalent to the removal of about 2.5 turns from a 60-base pair length of B-DNA. Untwisted, duplex DNA within the rec A-DNA complex remains base-paired (42). T antigen and arec A protein shareregions of homology over approximately 300 amino acids of the protein sequence (60). Perhaps the T antigen double hexamer surrounds the DNA duplex, and the DNA within isomerizes into an untwisted structure that is then bound and stabilized by the surrounding protein. The energy for the untwisting reaction must come from the protein-DNA contacts that are formed in the structure since ATP hydrolysis is not required. The Effect of ATP Hydrolysis and DTT on UntwistingWhen T antigen and DNA were incubated in the presence of ATP and the products examined by electron microscopy, many examples of what appeared to be hexamers of T antigen were observed bound at multiple sites along the DNA.' These hexamers apparently exert little untwisting effect, since in the presence of ATP untwisting is almost completely origindependent. However, in the presence of ATPyS, substantial origin-independent untwisting was observed. This effect is likely due to hexamers bound randomly to the DNA. It was observed in gel mobility shift experiments that complexes formed with T antigen and origin-containing DNA fragments in thepresence of ATPyS were much more stably bound than complexes formed in the presence of ATP (29).4Scheffner et al. (55) have also observed sequence-independent DNA binding by T antigen. In theabsence of DTT, therewas also substantial untwisting of ori- DNA. One possible reason for thiseffect is that in the absence of DTT, an improper disulfide bond forms within T antigen that increases the rigidity of the protein tertiary structure. Inthe case of bacteriophage T4 lysozyme, the mutational introduction of novel disulfide bonds can lead to increased protein stability, presumably by reducing the conformational fluctuations of the cross-linked chains (56). Thus, omission of DTT from an untwisting reaction or replacing ATP with ATPyS may have similar effects on the reaction. This could be true if in each case the rigidity of T antigen is increased to the extent that it can more easily assemble on a nonspecific DNA sequence and distort its helical structure. The effect of DTT may be to allow protein flexibility by inhibiting the formation of a disulfide bond. This would do two things. First, it could eliminate non-sequence-specific untwisting in thepresence of ATP, andsecond, it would allow a structural transition in T antigen to occur that is required for the initiation of DNA unwinding. The presence of DTT has no effect on ATP hydrolysis by T antigen, either in the presence or absence of DNA (data not shown). At present, it is unclear how two apparently unrelatedconditions, the omission of DTT and theuse of ATPyS, both result in the marked loss of specificity of T antigen for the SV40 origin. Acknowledgments-We thank Nilda Belgado and Barbara Phillips for technical assistance and David Valentin for preparation of artwork. E. coli DNA gyrase and E. coli topoisomerase I were gifts from Dr. Kenneth Marians. We thank Drs. KennethMarians, Russell

DiGate, Jim Borowiec, Peter Bullock, Stewart Shuman, and Samuel Rabkin for helpful discussions. REFERENCES 1. Li, J. J., and Kelly, T. J. (1984) Proc. Natl. Acad. Sci. U. S. A . 81,6973-6977 2. Stillman, B. W., and Gluzman, Y. (1985) Mol. Cell. Biol. 6,20512060 3. Wobbe, C. R.,Dean, F., Weissbach, L., and Hurwitz, J. (1985) Proc. Natl. Acad. Sei. U. S. A . 8 2 , 5710-5714 4. Murakami, Y., Wobbe, C. R., Weissbach, L., Dean, F. B., and Hurwitz, J. (1986) Proc. Natl. Acad. Sei. U. S. A . 8 3 , 28692873 5. Yang, L., Wold, M. S., Li, J. J., Kelly, T. J., and Liu, L. F. (1987) Proc. Natl. Acad. Sci. U. S. A . 8 4 , 950-954 6. Wobbe, C. R., Weissbach, L., Borowiec, J., Dean, F. B., Murakami, Y., Bullock, P., and Hurwitz, J. (1987) Proc. Natl. Acad. Sci. U. S. A . 8 4 , 1834-1838 7. Fairman, M. P., and Stillman, B. (1988) EMBO J. 7,1211-1218 8.. Wold, M. S., and Kelly, T. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,2523-2527 9. Prelich, G., Kostura, M., Marshak, D. R., Matthews, M. B., and Stillman, B. (1987) Nature 3 2 6 , 471-475 10. Lee, S.-H., Eki, T., and Hurwitz, J. (1989) Proc. Natl. Acad. Sci. U. S. A . 8 6 , 7361-7365 11. Weinberg, D. M., and Kelly, T. J. (1989) Proc. Natl. Acad. Sci. U. S. A . 86,9742-9746 12. Tsurimoto, T., and Stillman, B. (1989) Mol. Cell. Biol. 9 , 609619 13. Lee, S.-H., Kwong, A. D., Pan, Z.-Q., and Hurwitz, J. (1991) J. Bwl. Chem. 266,594-602 14. Ishimi, Y., Claude, A., Bullock, P., and Hunvitz, J. (1988) J. Biol. Chem. 263,19723-19733 15. Dean, F. B., Bullock, P., Murakami, Y., Wobbe, C. R., Weissbach, L., and Hurwitz, J. (1987) Proc. Natl. Acad. Sci. U. S. A . 84, 16-20 16. Wold, M. S., Li, J. J., and Kelly, T. J. (1987) Proc. Natl. Acad. Sci. U. S. A . 84,3643-3647 17. Tsurimoto, T., Fairman, M. P., and Stillman, B. (1989) Mol. Cell. Bwl. 9,3839-3849 18. Dodson, M., Dean, F. B., Bullock, P., Echols, H., and Hurwitz, J. (1987) Science 2 3 8 , 964-967 19. Kenny, M.K.,Lee, S.-H., and Hurwitz, J. (1989) Proc.Natl. Acad. Sei. U. S. A. 86,9757-9761 20. Brill, S. J., and Stillman, B. (1989) Nature 3 4 2 , 92-95 21. Deb, S., Tsui, S., Koff, A., DeLucia, A., Parsons, R., and Tegtmeyer, P. (1987) J. Virol. 6 1 , 2143-2149 22. Dean, F. B., Borowiec, J. A., Ishimi, Y., Deb, S., Tegtmeyer, P., and Hurwitz, J. (1987) Proc. Natl. Acad. Sci.U. S. A. 84,82678271 23. Bullock, P., Seo, Y. S., and Hurwitz, J. (1989) Proc. Natl. Acad. Sci. U. S. A . 86,3944-3948 24. Virshup, D. M., Kauffman, M. G., and Kelly, T. J. (1989) EMBO J. 8,3891-3898 25. Stahl, H., Droge, P., and Knippers, R. (1986) EMBO J. 5 , 19391944 26. Wiekowski, M., Schwarz, M. W., and Stahl, H. (1988) J. Biol. Chem. 263,436-442 27. Goetz, G. S., Dean, F. B., Hurwitz, J., and Matson, S. W. (1988) J. Biol. Chem. 263,383-392 28. Tjian, R. (1978) Cell 13, 165-179 29. Dean, F. B., Dodson, M., Echols, H., and Hunvitz, J. (1987) Proc. Natl. Acad. Sci. U. S. A . 84,8981-8985 30. Deb, S. P., and Tegtmeyer, P. (1987) J. Virol. 61,3649-3654 31. Borowiec, J. A., and Hurwitz, J. (1988) Proc. Natl. Acad. Sci. U. S. A. 8 5 , 64-68 32. Mastrangelo, I. A., Hough, P. V.C., Wall, J. S., Dodson, M., Dean. F. B.. and Hurwitz. J. (1989) Nature 338. 658-662 33. Borowiec, J. A , , and Hunvitz, J: (1988) EMBO J. 7,3149-3158 34. Parsons, R., Anderson, M. E., and Tegtmeyer, P. (1990) J. Virol. 64,509-518 35. Borowiec, J. A., Dean, F. B., Bullock, P. A., and Hurwitz, J. (1990) Cell 60,181-184 36. McClarin, J. A., Frederick, C. A., Wang, B.-C., Greene, P., Boyer, H. W., Grable, J., and Rosenberg, J. M. (1986) Science 2 3 4 , 1526-1541

DNA Untwisting by TAntigen 37. Saucier, J.-M., and Wang, J. C. (1972) Nature New Biol. 239, 167-170 Natl. 38. Wang, J. C., Jacobsen, J. H., and Saucier, J.-M. (1977) Nucleic Acids Res. 5, 1225-1241 39. Gamper, H. B., andHearst, J. E. (1982) Cell 29,81-90 40. Crick, F. H. C. (1976)Proc. Natl. Acad. Sci. U. S. A. 73, 26392643 41. Dodson, M., Roberts,J., McMacken, R., and Echols, H. (1985) 963 54. Proc. Natl. Acad. Sci. U. S. A. 82, 4678-4682 42. Di Capua, E., and Muller, B. (1987) EMBO J . 6, 2493-2498 43. Dean, F. B., Lee, S.-H., Kwong, A. D., Bullock, P., Borowiec, J . A,, Kenny, M. K., Seo, Y. S., Eki, T., Matsumoto, T., Hodgins, G., and Hurwitz, J . (1990) in Molecular Mechanisms in DNA Replication and Recombination (Richardson, C., and Lehman, R., eds) pp. 315-326, Alan R. Liss, Inc., New York 44. Giacherio, D., and Hager, L. P. (1980)J . Biol. Chem. 255,89638966 45. Myers, R. M., Kligman, M., andTjian, R. (1981) J . Biol. Chem. 256, 10156-10160 46. Roberts, J . M. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 39393943 47. O’Reilly, D. R., and Miller, L. K. (1988) J. Virol. 62, 3109-3119 48. Ryder, K., Vakalopoulou, E., Mertz, R., Mastrangelo, I., Hough, P., Tegtmeyer, P., and Fanning, E.(1985) Cell 42, 539-548 49. DeLucia, A. L., Deb, S., Partin, K., and Tegtmeyer, P. (1986) J. Viral. 57, 138-144 63.

5071

50. Higgins, N. P., Peebles, C. L., Sugino, A,, and Cozzarelli, N. R. Proc. (1978) Acad. Sci. U. S. A. 75, 1773-1777 51. Depew, R. E., and Wang, J . C. (1975) Proc. Natl. Acad. Sci. U. S. A. 72,4275-4279 52. Shure, M., Pulleyblank,D. E.,and Vinograd, J. (1977) Nucleic Acids Res. 4, 1183-1204 53. H. Scheffner, Stahl, and M., Knippers, R., (1989) Cell 57, 955-

55. 56. 57. 58. 59. 60. 61. 62.

King, K., Benkovic, S. J., and Modrich, P. (1989) J. Biol. Chem. 264,11807-11815 Sheffner, M., Wessel, R., and Stahl, H.(1989) Nucleic Acids Res. 17,93-106 Matsumura, M., Signor, G., and Matthews, B. W. (1989) Nature 342,291-293 Kim, R., Modrich, P., and Kim, S.-H. (1984) Nucleic Acids Res. 12,7285-7292 Kim, R., and Kim, S.-H.(1983) Cold Spring HarborSymp. Quant. Biol. 47,451-455 Williams, R. C., and Spengler, S. J. (1986) J. Mol. Biol. 187, 109-118 Seif, R. (1982) Mol. Cell. Biol. 2, 1463-1471 McEntee, K., Weinstock, G. M., and Lehman, I. R. (1981) J . Biol. Chem. 256,8835-8844 DiCapua, E., Engel, A., Stasiak, A,, and Koller, Th. (1982) J . Mol. Biol. 157, 87-103 Stasiak, A., and DiCapua, E. (1982) Nature 299, 185-186