cooperative assembly of the bovine papilloma virus e l and e2

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 21, Issue of July 25, pp. 15795-15803, 1993 Printed in U. S. A.

Cooperative Assembly of the Bovine Papilloma Virus E l and E2 Proteins on the Replication Origin Requires an Intact E2 Binding Site*

(Received for publication, February 9, 1993, and in revised form, April 13, 1993)

Monika LuskySj, Jerard Hurwitzll, and Yeon-Soo Seoll From the $Cornel1 University Graduate School of Medical Sciences, Hearst Microbiology Research Center, Department of Microbiolom New York, New York I0021 and the %Graduate Program in Molecular Biology, Sloan-Kettering Cancer Center, New York;New York 10021

Using quantitative gel retardation assays the prop- erties of the bovine papilloma virus (BPV) origin recognition protein E l and the effect of the viral E2 protein on the binding of E l to BPV origin DNA were examined. As reported previously (Seo, Y. S., Mueller, F., Lusky, M., Gibbs, E., Kim, H.-Y., Phillips, B. and J. Hurwitz (1993) Proc. Natl. Acad. Sei. U. S. A. 90, 2865-2869), the E 1 protein binds specifically to DNA sequences within the BPV origin (ori’) of replication. We also show that the presence of MgClz and ATP could stabilize the E 1 . ori+ DNA complex. At low levels of E 1, ori+ DNA binding was greatly stimulated by the viral E2 protein when the intact E2 binding site 1 2 was present on the DNA. In addition DNA-protein complexes formed in the presence of both E l and E2 were more stable than those formed with E l alone. In the absence of an E2 binding site the E2 protein inhibited the binding of E l to the BPV origin. Spacing of 0 or 9 base pairs between the E l binding site and the E2 binding site 1 2 abolished the stimulation of El-DNA binding by E2, whereas spacing of 6 base pairs between the two binding sites allowed for efficient stimulation. The data presented account for a direct role of E 2 in BPV DNA replication. We propose that the cooperative binding of both the E l and E2 proteins to BPV ori+ DNA is mediated by protein-protein interactions and by protein-DNA interactions, which include the formation of specific contacts of E2 with DNA.

Bovine papilloma virus (BPV)’ provides an interesting system to study the control of replication because the proviral DNA is maintained as a freely replicating plasmid at a con- stant copy number in BPV-transformed rodent cells (1). Only two viral proteins, E l and the viral transactivator E2, are required for BPV DNA replication i n vivo (2). I n vivo and in vitro studies (3,4) also identified a minimal BPV origin (ori+) sequence (nt 7911-22 of BPV DNA) that contains a binding site for El (El BS) centered around the unique HpaI site of the BPV DNA and half of an E2 binding site (E2 BS12). A cell-free in vitro replication system has been described (4) in

* This work was supported by Grants 5R0 CA51127-01 (to M. L.) and GM 34559 (to J. H.) 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 “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

6513; Fax: 212-746-8587. § T o whom correspondence should be addressed. Tel.: 212-746-

’ The abbreviations used are: BPV, bovine papilloma virus type I; SV40, simian virus 40; T antigen, SV40 encoded large tumor antigen; ori+, origin of replication; bp, base pair(s); BS, binding site; nt, nucleotide(s).

which DNA synthesis was absolutely dependent on the BPV minimal ori and on the E l protein. It was also reported that the E2 protein markedly stimulated the replication reaction in the presence of El, and at low levels of El , E2 was absolutely required (4).

Recently we described the overexpression and purification of the BPV E l protein (5) and confirmed that E l supported the synthesis of BPV ori+ DNA in vitro. These studies also showed that El shared a number of biochemical characteris- tics with the SV40 T antigen (6-8) including a DNA-dependent ATPase activity, DNA helicase activity, ori+ DNA binding activity, and the ability to unwind superhelical ori+ DNA, leading to the production of highly unwound DNA. These results clearly established E l as an initiator protein for BPV replication.

The role of the E2 transactivator protein in BPV replication is less clear. The E2 protein could indirectly affect BPV DNA synthesis by regulating the expression of viral and cellular genes required for replication. However, i n vivo studies showed that the absolute requirement for E2 remained even when the expression of the E l and E2 genes was driven from heterologous promoters and thus was uncoupled from each other (2). Moreover, it was shown that the transactivation function of E2 is not required for its role in replication (9). Together, these results indicate that E2 might play a more direct role in BPV replication. The full-length E2 protein can form a complex with the E l protein in solution (10, 11), suggesting that protein-protein interactions between E l and E2 might be important for some stages of BPV DNA synthesis.

In order to gain insight into the initial steps of BPV DNA synthesis we have investigated in more detail the binding of the El protein to the BPV origin DNA. Using gel retardation assays the influence of E2 on the binding of E l to ori+ DNA was examined. Furthermore, the effect of spacing between the E l BS and the E2 BS12 within the origin was probed. Our results and those described elsewhere (12) demonstrate that the presence of the complete E2 BS12, next to the E l BS, is crucial for the enhancement of E l binding and for the stabilization of E l at the origin by the E2 protein.

MATERIALS AND METHODS

E l and E2 Proteins-The generation of recombinant baculoviruses containing the wild type El and E2 genes were described previously (11) as was the purification of E l (5). The E2 protein was purified by DNA affinity chromatography (13) with the modifications described by Seo et al. (12).

Antisera-The use of polyclonal E l antisera (anti-El) was de- scribedpreviously ( l l ) , whereas polyclonal E2 antisera (anti-E2) were a generous gift from Dr. E. Androphy (Tufts University, Boston).

DNA Probes-The BPV sequences of the plasmid DNAs used in these studies are shown in Fig. 1. Plasmid pKSO (nt 7805-100; Ref.

15795

15796

FIG. 1. DNA sequence of the BPV origin region and plasmid con- structs containing wild type and mutant sequences. A, BPV DNA sequence including the BPV origin and the E2 BSll and BS12. The minimal origin (4) is indicated by brackets. The A/T- rich region is contained within the gray stippled box, whereas the E l BS is contained within the black box, indicated above the sequence. E2 BSll and E2 BS12 are indicated by hatched boxes. Arrows indicate the inverted repeat IR1 (3) within the E l BS. B, schematic rep- resentation of plasmids containing wild type and altered BPV sequences. The construction of the plasmids is described under "Materials and Methods." The small white box represents the three hp CAC (nt 13-15) between the E l BS and E2 BS12. pUCOM-BS12-0 has these 3 bp deleted, whereas pUCOM-BS12-6 and pUCOM-BS12-9 contain 6 bp (CACCAC) and 9 bp (CACCACCAC) between the E l BS and E2 BS12, respectively. pUCO-Xho contains a XhoI linker (X) in the center of the E l BS at the HpaI site (nt 1).

Origin Recognition by the BPV E l and E2 Proteins

minimal origin

I 1 E2 BSl l A/T rich El BS E2 BS12

"4 A C C G A A A C C G C T A A G T R A A G A C T A T G T A m T T T C C C A G T G A .

0

7900 . 7920

0 0

7940 1

BS11 A / T rich El BS BS12

4) containing the BPV origin, including the complete E2 BSl l and BS12, was a gift from Dr. M. Botchan (University of California Berkeley). Plasmid pUCO-Xho, generated by insertion of a XhoI linker into the unique HpaI site (nt 1) of the BPV origin in pUC/ Msp (nt 7903-81; (3)) was a gift of Dr. A. Stenlund (Cold Spring Harbor, NY). Plasmid pUCOM (nt 7911-221, described previously (5), contains the BPV origin and half of the E2 BS12. Plasmid pUCOMABS12 (nt 7911-16) contains the minimal BPV origin without any E2 BS cloned into the BamHI site of pUC19. Plasmids pUCOM-BS12-0, pUCOM-BS12-6, and pUCOM-BS12-9 were derived from pUCOMABS12. Briefly, the sequences between the BPV HpaI site (nt 1) and the XbaI site in pUC19 were deleted from pUCOMABS12 and replaced by synthetic oligonucleotides containing the sequences shown in Fig. 1. pUCOM-BS12-0 has 3 bp (CAC, nt 13-15) deleted, whereas pUCOM-BS12-6 and pUCOM-BS12-9 contain 6 (CACCAC) and 9 (CACCACCAC) bp, respectively, between the E l and E2 binding sites. Similarly, pUCOM-BS10 was generated by replacing the sequences between the BPV HpaI site and the XbaI site in pUCOMABS12 with a synthetic oligonucleotide containing the E2 BS10. Linear duplex DNA fragments used to assay the binding activities of the E l and E2 proteins were prepared by isolating the 232-bp EcoRI-BamHI restriction fragment of pKSO and the EcoRI-

. 20

pKSO (wt)

pUCOM

PUCO-Xho

pUCOMb BS12

pUCOM - BSl2-0

pUCOM - BS12-6

BS10

pUCOM - BS10

protein, as indicated, were assembled on ice and incubated for 15 min at 37 "C. Reactions were terminated by the addition of glutaraldehyde (0.2% final concentration) and were further incubated for 15 min at the reaction temperature. For supershift assays, complete reactions were assembled as above and incubated for 5 min at 37 "C. Upon addition of antibodies, as indicated, reactions were incubated for additional 10 min and were terminated as above. The formation of DNA-protein complexes was analyzed by electrophoresis through 1.8% horizontal agarose gels in 0.5 X TBE (45 mM Tris, 45 mM boric acid, 1 mM EDTA) for 6-7 h at 65 V. Gels were dried and subjected to autoradiography.

Dissociation Rates-Binding reactions were scaled up %fold with the indicated amounts of E l in the absence or presence of E2 protein. After incubation for 7 min at 37 "C, at which time binding had reached equilibrium (data not shown), a 20-p1 aliquot was removed (time 0) and a 50-fold molar excess of unlabeled pKSO DNA (1.5 pmol/time point) was added immediately to the reactions. Aliquots (20 pl) were removed from the reactions at subsequent time points, as indicated, transferred into tubes containing glutaraldehyde, and DNA-protein complex formation was analyzed as described above.

RESULTS

Origin Recognition by the BPV El and E2 Proteins 15797

Bound DNA

Free DNA -

I 2 3 4 5 6 7 R 9 1 0 1 1 1 2

0.

P .L /

Bound DNA

Free DNA , 2 3 4 5 6 7 8

Complex 2 2 2 0 2 formed (fmol)

FIG. 2. Requirements for E l binding to wild type BPV ori+ DNA measured by gel retardation. Reaction mixtures containing 30 fmol of the :'ZP-labeled EcoRI-RamHI fragment of pKSO DNA were assembled and processed as described under "Materials and Methods." A, influence of MgClz and ATP on the binding of El . Increasing amounts of El were incubated a t 37 "C for 15 min with the ori' DNA (pKSO) fragment under the conditions indicated above the lanes. After incubation, reactions were terminated with glutaraldehyde and DNA-protein complex formation was analyzed by electrophoresis through a 1.8% horizontal agarose gel in 0.5 X TBE. The gel was dried and subjected to autoradiography. Arrows on the left of the autoradiogram indicate the positions of free and El-complexed (bound) DNA in the gel. E, quantitation of the complex formation shown in A. Complex formation was measured (expressed in femtomoles) in the presence of M$Iz alone (A) or in the presence of both M$12 and ATP (0) or in the absence of both (0). Background values, which were normally between 1 and 2% of the input substrate, have been subtracted from the values presented. C, specificity of E l binding to ori' DNA. DNA binding reactions containing 100 ng of E l protein were assembled and processed as described above with the omission or addition of the indicated components: lane I , El protein was omitted; lane 2, M g C I 2 and ATP omitted; lane 3, ATP omitted; lane 4, complete reactions containing 7 mM MgC12 plus 4 mM ATP; in all other lanes complete reactions were performed lane 5, omission of glutaraldehyde; lane 6, omission of nonspecific competitor DNA; lane 7, omission of nonspecific competitor DNA and the addition of 600 fmol (20-fold molar excess) of unlabeled ori' (pKSO) DNA; lane 8,

worth noting. ( a ) Specific binding was observed in the absence of MgClz and ATP (Fig. 2 A , lanes 2-4), which was reduced by addition of MgC12 (Fig. 2 A , lanes 6-8, see also Fig. 2C, lane 3). ( 6 ) In the presence of MgC12, ATP addition stimulated E l binding 3-fold at the highest level of E l used (Fig. 2 A , lane 12 ) . ( c ) Under all three conditions the amount of complex formed showed a sigmoidal response to the concentration of E l added, suggesting that the E l protein forms multimers at the BPV origin. ( d ) The complex formed in the presence of both MgC12 and ATP resulted in the retardation of a more distinct DNA-protein species compared with the heteroge- neous population of complexes observed in the absence of MgCl, and ATP (see also Fig. 3). Fig. 2C demonstrates that in the absence of glutaraldehyde a distinct DNA-protein complex was not detected (lane 5). Glutaraldehyde addition was also required when reactions were carried out in the absence of MgClz and ATP (data not shown). Fig. 2C, lanes 7 and 8 demonstrate that the binding of E l t o labeled ori+ DNA was completely abolished in the presence of a 20-fold molar excess of cold ori+ DNA (lane 7) but not in the presence of unlabeled ori- DNA (lane 8) . The same results were obtained in the absence of MgC12 and ATP (data not shown).

Dissociation Rate of El in the Absence and Presence of ATP-The rate of dissociation of E l from ori+ DNA (pKSO) was determined in terms of the half-life of E l .DNA complex formation. A 50-fold molar excess of unlabeled ori+ (pKSO) competitor DNA was added to binding reactions that had reached equilibrium (Fig. 3). The ratio of bound DNA at the indicated time points to bound DNA at time 0 was measured. The data are plotted in Fig. 3i3 as the log of the complex (counts/min) present a t time 0 divided by the complex (counts/min) present a t various times after the addition of competitor DNA. The results show that the half-life of E l in the presence of MgC12 and ATP is increased 2-fold over that observed in the absence of MgC12 and ATP.

Effect of E2 on the Binding of E l to ori+ (PKSO) DNA- DNAse I footprinting studies have shown that the E2 protein enhanced the binding of E l t o an 18-bp palindromic sequence centered around the unique HpaI site in BPV DNA (see Fig. 1) approximately 10-fold when the E2 binding sites BSll and BS12 were present on the DNA (4). Fig. 4 shows that E l .o r7 DNA complex formation with pKSO DNA was stimulated in the presence of E2 both in the absence and presence of MgCl2 and ATP. However, the stimulation was greater in the presence of both components. The largest extent of stimulation (30-40-fold) was observed at low levels of E l (Fig 4, A and B, lanes 1 1 and 12). Increasing amounts of E l in this assay reduced the stimulation effect to between 3- and 4-fold (Fig. 4, A and B, lane 14). The data presented also show that the DNA-protein complexes formed in the presence of E l plus E2 migrated on the gels as broad bands and included species that migrated faster than those obtained with E l alone (see also Figs. 6 and 8). We did not observe DNA-protein complexes that migrated significantly slower than those obtained with El alone.

Under these assay conditions, stable E2 binding to B S l l and BS12 contained within pKS0 DNA, in the absence of El, was not observed above background levels (Fig. 4, A and B, lanes 8 ) with the level of E2 used (60 ng). Using the same assay conditions, binding of E2 to BSlO was observed (see Fig. 11). It was reported previously that E2 binding to BSll and BS12 was 11- and 100-fold lower than the binding to BSlO (15). In our studies we observed weak E2 binding to

omission of nonspecific competitor DNA and the addition of 600 fmol of ori- (pUC19) DNA. Numbers at the bottom indicate the amount of complex formation.

15798 Origin Recognition by the BPV E l and E2 Proteins

Bound DNA - ____) " FreeDNA 1 2 3 4 5 1 2 3 4 5

' I

0 2 4 6 8 10 12 I4 16 18 20

minutes aller addition of competitor DNA

FIG. 3. Effect of MgClz and ATP on the stability of the E l . ori+ DNA complex. Binding reactions were increased 8-fold and contained a total of 240 fmol of "P-labeled pKSO DNA fragment and a total of 800 ng of El protein (100 ng for each individual time point). A, gel retardation. The rate of dissociation of El. DNA complexes in the absence or in the presence of MgC12 and ATP, as indicated, was monitored at the indicated times after addition of a 50-fold molar excess of unlabeled pKSO DNA (1.5 pmol for each individual time point) to the binding reactions after 7 min of incubation (time 0). B, quantitation of the data shown in A. Ratio of labeled DNA shifted (count/min) at time x over the amount of labeled DNA shifted (count/min) a t time 0 are plotted uersw min after the addition of competitor DNA on a semi-log graph. The half-lives of the complexes derived from this plot are: 6 min in the absence of MgCI, and ATP (W) and 13.5 min in the presence of both MgC1, and ATP (0).

BSll and BS12 only in McKay assays (data not shown). To ascertain that the stimulatory effect was due to both

the E l and E2 proteins present in the DNA-protein complexes, El- and E2-specific antibodies were included in the binding reactions, and supershift assays were performed. Ad- dition of either antibody to DNA in the absence of proteins did not alter the migration of free DNA (Fig. 5, lanes 9 and 15) . However, the addition of anti-El to reactions containing either E l alone (lanes IO and 11 ) or E l plus E2 (lanes 12 and 13) further retarded the migration of the DNA-protein complexes. In the presence of E2 antibodies, the migration and amount of E l . DNA complex formed in the absence of E2 was not changed (lunes 16 and 17), confirming the specificity of the antibody preparation. However, the addition of anti-E2 to reactions containing both E l and E2 proteins resulted in a supershift of DNA-protein complexes, indicating the presence of E2 in the respective complexes. In all reactions where the addition of either E l or E2 antibodies resulted in supershifted

Bound DNP - Free DNA ___)

Bound DN, -

C.

c E - 20 '0 e, !E 0 X e,

c

- E lo

V

0 0 20 40 60 80 100 120

E l added (ng)

FIG. 4. Influence of the BPV E2 protein on the binding of E l to ori+ (pKSO) DNA in the absence ( A ) and in the presence ( B ) of MgCl, and ATP. Increasing amounts of El protein either in the absence (-) or presence (+) of 60 ng of E2 were incubated with the R2P-labeled pKSO DNA fragment and reactions were assembled and processed as described under "Materials and Methods" and in the legend to Fig. 3. C, quantitation of the data from A and R. Binding was measured as the amount of complex formed (femtomoles) in the absence (A, A) or presence (0, 0) of both M$I* and ATP, either with (A, 0) or without (A, 0) the E2 protein.

DNA-protein complexes, the amount of complex formed was also increased. We do not believe that this activation is specific as recently shown for the activation of p53 binding by certain monoclonal antibodies (16), since the antibodies used here were polyclonal. It is likely that in our case the antibodies stabilized the complexes during the electrophoretic separation and thus increased the amount of complex formed.

Taken together, the results presented in Figs. 4 and 5 suggest that the interplay between E l and E2 mediates cooperative binding of E l to the BPV origin DNA in the presence of E2 BSll and E2 BS12.

Origin Recognition by the BPV El and E2 Proteins 15799

- ab + anti-El + anti-E2 "- E2 (60ng) - - - - + + + + - - - + + + - " + + +

El (ng) o:ZO:ZOo o Z O ~ 8 0 0 Z O Z O o 0 0 0 0 0

.~ , """"- -- - "

Bound DNA -

1 2 3 4 5 6 7 8 9 10 1 1 1 2 1 3 1 4 1 5 16 1 7 1 8 1 9 2 0

Complex 2 - '9 m formed h m m m v! 2 m

N v! m y !

(fmol) * Z P

FIG. 5. Effect of E l and E2 antibodies on the El-dependent ori* DNA complex formation in the presence and absence of E2. Increasing amounts of El without (-) or with (+) E2 protein (60 ng) were incubated with the '*P-labeled pKSO DNA fragment under the conditions described under "Materials and Methods." All reactions contained MgCI, and ATP. In the panels marked +anti E2 and +anti E2, 1 pg of the polyclonal antisera, respectively, was added after 5 min of incubation; reactions were further incubated for 10 min and terminated with glutaraldehyde. The stippled arrow at the left of the autoradiogram indicates the position of supershifted DNA- protein complexes in the gel. The amount of complex formed (femtomoles) are presented at the bottom of the figure.

Dissociation Rate of El in the Presence of E2"It was of interest to determine whether the presence of E2 affected the stability and thus influenced the half-life of the E l . ori+ DNA complex. For this reason, challenge experiments, with unlabeled ori+ DNA, were performed (Fig. 6). The reactions contained 50 ng of E l either in the absence or presence of 60 ng of E2 for each individual time point and were carried out in the presence of MgCl, and ATP. Quantitation of these data (Fig. 6B) showed that the half-life of the DNA-protein complex formed in the presence of E l plus E2 was extended approximately 2-fold over the half-life of the complex formed with E l alone. These results suggest that E2 not only enhances E l binding to ori+ DNA but also stabilizes the DNA-protein complex.

Stimulation of El Binding by E2 Requires the Presence of the Intact E2 BS12"Previous reports on the physical association between the E l and E2 proteins (10, 11) suggested that the interaction between E l and E2 might play a role in the recognition of the origin by El . Because of this, we determined whether the presence of an E2 BS (and thus E2 binding) was required for the stimulatory effect of E2 on the binding of E l to ori+ DNA. The binding of E l in the absence and presence of E2 was monitored using the DNA substrate pUCOMABS12, which contains the BPV sequences nt 7911- 16, including the E l BS, but has both the E2 BSll and BS12 deleted (Fig. 1). Fig. 7 demonstrates that BPV sequences beyond nt 16 were not required for efficient El binding (lanes 2-4). However, the presence of E2 did not stimulate E l binding to pUCOMABS12 DNA (lanes 5-7). The addition of El antibodies resulted in a supershift of the complex, both in the absence and presence of E2 in reactions containing El (lanes 10-13). The addition of E2 antibodies did not alter the mobility of the DNA-protein complexes in reactions where El plus E2 were present (lanes 18 and 19). We conclude from these results that El-E2 protein interactions alone are not

Bound DNA ___)

Free DNA -

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4

: ~ 1 1 1 1 1 , 1 1 1 ~

4

0 4 8 12 16 20 24 28 32 36 40

minules atler addilion 01 competitor DNA

FIG. 6. Rate of dissociation of E l containing DNA-protein complexes in the absence and presence of the E2 protein. A, binding reactions containing MgCI, and ATP were assembled and processed as described under "Materials and Methods" and in the legend to Fig. 3. Reactions contained 50 ng of El for each individual time point either in the absence (lanes 2-7) or presence (lanes 8-14) of E2 protein (60 ng for each time point). The rate of dissociation was monitored at the indicated times upon addition of a 50-fold molar excess of unlabeled pKSO DNA per time point. E , data from A are plotted on a semi-log graph as described in the legend to Fig. 3. The half-life from this plot are: 11 min in the absence of E2 (0); 26.5 min in the presence of E2 (A).

sufficient to account for the stimulatory effect observed with pKSO DNA.

The results shown in Fig. 7 (lanes 2-7) also suggest that in the absence of an E2 BS the binding of E l was inhibited by E2. This was examined by adding increasing amounts of E2 to reactions that contained 50 ng of E l and comparing complex formation with pKSO DNA and with pUCOMABS12. Fig. 8 shows that with pKSO DNA El-dependent DNA- protein complex formation increased with increasing amounts of E2 (lanes 3-5). In contrast, using pUCOMABS12 DNA, the amount of DNA-protein complex was reduced by increasing amounts of E2 (lanes 9-13). At the highest levels of E2 used (60 and 120 ng), the amount of complex formed was reduced &fold. These results confirm those described in Fig. 7 and indicate that E2 inhibits E l binding in the absence of an E2 BS. One possible interpretation is that at high levels of E2 complex formation between E l and E2 is favored over complex formation between E l and DNA and that a preformed E l . E2 protein complex is not active in binding to the E l binding site in the absence of an E2 BS.

15800 Origin Recognition by the BPV El and E2 Proteins

- ab + anti-El + anti-E2 ml--"-l-

E2 (60 ng) - - - - + + + + - - - + + + - - - + + +

El(ng) , :S:$:S?, , S O O S , O S : $ S ? O 0 0

"_ -+ - Bound DNA -

Free DNA _i

1 2 3 4 5 6 7 S 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0

Complex '9 y! v! o? " Z N m 2 2 2 x formed O - O O

(fmol) -

FIG. 7. Lack of stimulation of El binding to the El BS in the absence of an E 2 BS. Increasing amounts of El, in the absence (-) or presence (+) of 60 ng E2 protein, with or without antibodies (as indicated), were incubated with 30 fmol of 32P-labeled pUCOMABS12 DNA fragment (see Fig. 1). Reactions containing MgCI, and ATP were assembled and processed as described under "Materials and Methods" and in the legend to Fig. 5. The amounts of complex formed (femtomoles) are presented at the bottom of the figure.

A. E2 (ng) El (50 ng) - + + + + + - - + + + + + -

. . "

The effect of E2 on E l binding was also examined using pUCOM DNA (nt 7911-22) containing the entire E l BS and half of the E2 BS12 (see Fig. 1). Under the conditions described in the experiments shown in Figs. 6 and 8, E2 stimulated E l binding to pUCOM DNA only 2-fold (data not shown; Ref. 12). Together, the results obtained with pUCOMABS12 and pUCOM DNAs suggest that for the cooperative binding observed with E l plus E2 on pKSO DNA, binding of E2 to either E2 BSll or BS12 is required.

In uiuo (3) and in uitro (4) studies indicated that E2 BSll was dispensable for origin function. Thus, the stimulatory effect of E2 on E l binding to wild type pKSO DNA is most likely mediated through protein-DNA interactions which in- volve E2 BS12 (see below).

The binding of E l in the absence and presence of E2 was also measured using an ori--DNA substrate, pUCO-Xho. This DNA contains an intact E2 BS12 and a XhoI linker inserted in the center of the E l BS (Fig. 1). In uiuo studies have shown that this DNA template does not support DNA replication in the presence of E l and E2 (3). In addition, a similar linker insertion did not support El- and E2-dependent DNA synthesis in uitro (4). Fig. 9 shows that hardly any complex formation occurred with increasing amounts of E l in the absence of E2 (lanes 2-4). However, E2 addition stimulated the complex formation (lanes 5-7), although the amount of complex formed was considerably lower than that observed with pKSO DNA. The data presented in Fig. 9 establish that the presence of the intact E2 BS12 is crucial for the stimulatory effect of E2 on E l binding to ori+ DNA and for the stability of ori+ DNA-protein complexes.

Influence of Spacing between the El BS and E2 BS12 on

Bound DNA

Free DNA

B.

"

1 2 3 4 5 6 7 8 9 10 11 1213 14

0 50 100

E2 added (ng) 150

FIG. 8. E2 inhibits the binding of El to BPV ori* DNA in the absence of an E 2 BS. A , 50 ng of E l and increasing amounts of E2 (as indicated) were incubated with labeled pKSO DNA (lanes 1-7) or pUCOMABS12 DNA fragments (lanes 8-14) and the formation of DNA-protein complexes was analyzed. B, quantitation of the data shown in A. Complex formation (expressed in femtomoles) was measured with increasing amounts of E2 using pKSO DNA (0) or pUCOM BS12 DNA (0) as substrate.

Bound DNA - -

Free DNA 1 2 3 4 5 6 7 8

' t / oo- 100 200 300

El added (ng)

FIG. 9. Influence of E2 on the binding of El to BPV ori- DNA. A , increasing amounts of E l in the absence (-) or presence (+) of 60 ng of E2 protein were incubated with the 32P-labeled pUC0- Xho DNA fragment (see Fig. 1). Complex formation was analyzed as described. B, quantitation of the data shown in A . Complex formation (expressed in femtomoles) was measured in the absence (0) or presence (0) of E2.

Origin Recognition by the

the Effect of E2 on El Binding-To gain further insight into the mechanisms underlying the enhancement effect of E2, we asked whether the distance between the two binding sites, ElBS and E2 BS12, was important. Three mutant substrates were generated and compared with pKSO DNA. pUCOM- BS12-0 has the 3 bp (CAC) (nt 13-15) between the ElBS and E2BS12 deleted, whereas pUCOM-BS12-6 and pUCOM- BS12-9 contain 6 and 9 bp, respectively, between the E l BS and the E2 BS12 (see Fig. 1). The binding of E l in the absence and presence of E2, as well as in the absence and presence of El and E2 antibodies was examined. The results are presented in Fig. 10. E l alone interacted with all three substrates with equal efficiency (Fig. 10, A and B, lanes 2 4 , and data not shown), indicating that the 3 bp 3' to the "El palindrome" did not contribute to the E l BS. The presence of E2 neither stimulated nor decreased the amount of E l . DNA complex formed when pUCOM-BS12-0 (Fig. lOA, lanes 5-7) and pUCOM-BS12-9 (data not shown) were used as substrates. Addition of E l antibodies supershifted the DNA-protein complexes formed both in the absence and presence of E2 (Fig. 10 A and B, lanes 10-13). However, the addition of E2 antibodies had no effect on the mobility of the DNA-protein complexes (Fig. lOA, lanes 18 and 19). We conclude from these results that deletion of the 3 bp between the E l BS and the E2 BS12 (pUCOM-BS12-0) or extending the distance to 9 bp (pUCOM-BS12-9) interfered with efficient El-E2 protein interactions and protein DNA interactions.

Surprisingly, the results obtained with pUCOM-BS12-6 DNA (Fig. 10B) which contains a total of 6 bp (CACCAC)

A. - ab + anti-El + anli-E2

E2(60flg) - . - - + + + + - - - + + + - - - + + + "m ~ ~ ( n g ) , C O E P O E , , 0 ! 3 ~ 8 , , s E s X ,

."" _ _ _ _ __. . -

"", -

"")

Bound DNA

Free " DNA 1 2 3 4 5 6 7 8 3 10 11l?l3 14 1516 171819 LU

FIG. 10. Effect of distance between the ElBS and E2BS12 on the binding of E l to ori+ DNA in the absence and presence of E2. Increasing amounts of El without (-) or with (+) E2 protein (60 ng) were incubated with the 32P-labeled DNA fragments of pUCOM-BS12-0 ( A ) and pUCOM-BS12-6 ( B ) . Reactions were assembled and processed as described under "Materials and Methods." Supershift experiments with anti-El and anti-E2 were as described in the legend to Fig. 6. The amounts of complex formed (femtomoles) are presented at the bottom of the autoradiograms.

BPV El and E2 Proteins 15801

between the E l and E2 binding sites were similar to those obtained with wild type pKSO DNA (Fig. 5). Maximal stimulation of DNA-protein complex formation (16-fold) in the presence of E2 was observed at low levels of E l (25 ng), whereas the stimulation decreased (2.3-fold) at the highest amount of E l used (100 ng). The addition of E2 antibodies clearly caused a supershift in the DNA-protein complex when both E l and E2 were present (lanes 18 and 19), whereas the migration of complexes formed with E l alone was unaffected by anti-E2 (lanes 16 and 17). These data, together with those shown in Fig. 5, demonstrate that efficient El-E2 protein and protein-DNA interactions occur at a spacing of either 3 bp (pKSO) or 6 bp (pUCOM-BS12-6) between the two binding sites, but not when the binding sites are too close together (pUCOM-BS12-0) or too far apart (pUCOM-BS12-9). Exper- iments to test whether the separation of the two binding sites by 1 or 2 and 7 or 8 bp influences the stimulatory role of E2 on E l binding are currently under investigation. When the E2 BS12 was placed more distal, e.g. 30 bp 3' or 35 bp 5' to the E l BS, E2 had no effect on E l binding (data not shown).

Effect of the Strength of the E2 BS on the Influence of E2 on El Binding-It was of interest to ask whether the strength of the E2 BS next to the E l BS in wild type BPV ori+ DNA was important. For this purpose E2 BS12 was replaced with the strong E2 BSlO (15) generating pUCOM-BS10 DNA (Fig. 1). With this substrate, significant binding with E2 alone was observed (Fig. 11, lane 8). This result shows that under the conditions described here for binding and gel retardation E2 binding to a strong binding site such as BSlO can indeed be monitored; however, these conditions do not lead to stable binding of E2 to the weak E2 BS12.

The addition of E2 to reactions containing E l resulted in a marked stimulation of DNA-protein complex formation (Fig. 11, lanes 5-7). However, when the amount of complex formed with E2 alone was subtracted from the amount of complex formed with E l plus E2, it was apparent that complex formation was proportional to the amount of E l added. Thus a strong E2 binding site, such as BS10, used in place of the weak E2 BS12, resulted in a stimulation of E l binding by E2 but did not give rise to the sigmoidal response observed with wild type BPV ori+ DNA. Similar results were obtained when BSlO was placed 30 bp distal to the E l BS (data not shown).

DISCUSSION

Using quantitative gel retardation assays, we have demonstrated that the BPV E l protein specifically recognizes and binds to BPV ori+ DNA, but not to ori- DNA (Figs. 2 and 9). Efficient E l . ori+ DNA complex formation was observed in the absence and presence of MgC12 and ATP (Figs. 2 and 4). Interestingly, the amount of complex formed showed a sigmoidal response to the amount of E l added, regardless whether the binding was carried out in the absence or presence of MgC1, and ATP or in the presence of MgCl, alone. This cooperative binding suggests that E l forms multimers at the two half-sites of the BPV origin.

The ability of E l to bind ATP (17, 18) appears to be important for its replication functions, since a mutant in the ATP binding domain was shown to be replication defective in uiuo (11, 17) and in vitro.' Thus, ATP binding to E l might induce conformational changes or, alternatively, might mediate the formation of multimeric structures of defined size, analogous to what has been observed with SV40 T antigen (19, 20). In addition ATP-induced structural changes of the E l protein could affect the stability of El-ori' DNA com-

* M. Lusky, unpublished observations.

15802 Origin Recognition by the BPV El and E2 Proteins

A. E2 (60 ng) - - - - + + + +

El (ng) S' 2: : 0 0

. . . " - . - .

FIG. 11. Effect of E2 on the binding to BPV DNA containing the E2 BSlO in place of BS12. A , increasing amounts of E l without (-) or with (+) E2 protein (60 ng) were incubated with the 32P-labeled DNA fragment of pUCOM-BS10. Reactions were processed essentially as described in the previous figures. E, quantitation of the data shown in A . Complex formation (expressed in femtomoles) was measured in the absence (0) or presence (A) of E2 protein. The amount of complex formed in the presence of E2 alone was subtracted from the amount of complex formed in the presence of both proteins and is also indicated (0).

plexes. In support of this we observed that complexes formed in the presence of MgC12 and ATP were more stable (2-3- fold) than complexes formed in the absence of both components (Fig. 3). However, the precise nature of the effect of ATP on the structure of E l in the absence and presence of ori+ DNA needs to be determined. In uiuo studies have shown that in addition to E l the BPV

E2 transactivator protein was absolutely required for viral plasmid DNA replication (2). However, the precise role of E2 in replication is not clear. The finding that E l and E2 can form a protein complex in solution (10, 11, 21) suggests that the physical association between the two proteins might be important for some stages of BPV DNA synthesis.

Using wild type BPV ori+ DNA (pKSO) we showed that E l mediated origin binding can be dramatically stimulated in the presence of E2. Furthermore, the antibody supershift experiments, described in Fig. 5, suggest that both E l and E2 proteins were present in the DNA-protein complexes formed. In addition, the half-life of the protein-DNA complexes was increased 2-%fold in the presence of E2 compared with complexes that contained only E l (Fig. 6). We conclude that (i) the BPV E l protein plays a key role in origin recognition, and (ii) El-mediated origin recognition is stimulated and stabilized by the BPV E2 protein. We also reported that E2 can stimulate the El-specific unwinding of wild type ori+ DNA (12). Together, these data account for a direct role of E2 in replication.

Interestingly, the stimulatory effect of E2 on the binding of E l to wild type ori+ DNA (pKSO) was not absolutely dependent on the presence of MgC12 and ATP. This raises the question whether E2, in binding to El, might mediate conformational changes in E l analogous to those formed in the presence of ATP.

In contrast to the effect observed with pKSO, the stimulation of E l binding by E2 was small (2-fold) when only half of the E2 BS12 was present on the or? DNA (pUCOM; Ref. 12 and data not shown), and no stimulation was observed (Fig. 7) when DNA-protein complex formation was monitored with BPV ori+ DNA lacking the E2 binding sites BSll and BS12 (pUCOMABS12). Moreover, increasing amounts of E2 seemed to interfere with the binding of E l to this DNA, but not to wild type pKSO DNA (Fig. 8). These results indicate that maximal E2-mediated stimulation of E l binding to the BPV origin requires the presence of an intact E2 BS. We conclude from these results that E1-E2 protein-protein interactions alone are not sufficient for the E2 enhancement of El-or i recognition. In fact, the results presented in Fig. 8 suggest that formation of a productive E l E2 protein complex on the origin DNA is regulated by and requires specific contacts of both proteins with the DNA. Furthermore, these data are consistent with the idea that a preformed EleE2 protein complex might be in a nonproductive conformation which prevents or interferes with origin recognition at least in the absence of an E2 binding site. This notion is not without precedent; it was shown that preformed hexameric structures of SV40 T antigen rendered the protein inactive for binding to the site I1 core origin (19).

Interestingly, the results shown in Fig. 9, together with those described previously (12), indicate that E l .DNA complex formation can be stimulated by E2 in the presence of E2 BS12 and one half-site of the E l palindrome (pUCO-Xho), whereas E l binding or E2 binding alone to this DNA was not observed. These results further indicate that E2 plays a direct role in targeting E l to its binding site. We have speculated that E l might be assembled on two half-sites of the E l palindrome, and one multimer interacting with E2 complexed with BS12 may facilitate the assembly of the other multimer on the "far" side of the E l palindrome (12).

Taken together, our results imply that E2 binding to BS12, located next to the E l BS, is required for the stimulation and stabilization of E l origin recognition and binding. Our data are entirely consistent with the requirements observed for the in uiuo replication of BPV in transient transfection assays (3). Those studies (3) also reported that a deletion of the most carboxyl-terminal 33 amino acids, including the DNA binding and dimerization domains of E2, rendered the resulting E2 mutant defective in supporting BPV replication in uiuo. In addition, E2 mutants only affecting the DNA binding, but not the dimerization domain, failed to support BPV DNA synthesis, thus indicating that the DNA binding activity of E2 is required for replication (3). In a different study (9), however, a more complex set of phenotypes was observed with carboxyl-terminally truncated E2 mutants. In those studies (9) the same mutation as described by Ustav et al. (Ref. 3; deletion of 33 amino acids at the carboxyl terminus) also failed to support BPV replication. In contrast, certain E2 mutants missing more than 33 amino acids from the carboxyl terminus could support replication although a t reduced levels. I t is possible that the latter mutations lead to certain conformational changes in the E2 protein that could influence protein-protein and or protein-DNA interactions. Thus, it will be of interest to determine the effect of these E2 mutants on the binding of E l to BPV ori DNA in the absence and presence of E2 BS12.

Mutant BPV origin DNA substrates were used to further investigate the influence of E2 on E l binding. The results presented in Fig. 10 showed that the spacing between the E l BS and E2 BS12 within the BPV origin appears crucial for the stimulatory effect of E2. Deletion of the 3 bp (nt 13-15)

Origin Recognition by the BPV E l and E2 Proteins 15803

between the ElBS and E2 BS12 (Fig. 1OA) or insertion of a total of 9 bp (data not shown) between the two binding sites completely repressed the stimulatory function of E2. These data were further substantiated by the results that the DNA- protein complexes formed in the presence of E l and E2 were not supershifted by E2 antibodies. This indicates that E2 was not stably engaged in active E l containing DNA-protein complexes.

Surprisingly, when the E l and E2 binding sites were spaced apart from each other by a total of 6 bp (pUCOM-BS12-6) E2 stimulated the binding of E l (Fig. 10B) as well as it stimulated the binding of E l to wild type BPV ori’ DNA (pKSO). The antibody supershift experiments demonstrated that E2 was part of the DNA-protein complexes formed in the presence of both the E l and E2 proteins. Together, these data indicate that a rather precise spacing is required between the two binding sites in order to engage the E l and E2 proteins into a complex which results in cooperative binding.

Interestingly, it seems important that the E2 BS12 within the BPV origin is one of the weak E2 binding sites. As shown in Fig. 11 the E2 stimulation of E l binding to BPV ori+ DNA which contained the E2 BSlO instead of BS12 at the wild type position (pUCOM-BS10) did not reflect cooperative binding. It appeared that the E l and E2 proteins bound independently and did not communicate with each other.

The data described in this study support a direct role of E2 in BPV DNA replication and indicate that El-E2 protein- protein and El-E2/DNA interactions are a critical feature in the initial stages of BPV DNA synthesis. It will be important to investigate the subunit structure of E l .E2 protein complexes in the absence and in the presence of BPV origin-

containing DNA to gain insight into the mechanism of assembly of the origin recognition complex.

Acknowledgments-We thank Dr. F. Mueller and E. Gibbs for generously providing us with preparations of E2 protein. We thank Drs. M. Linden and K. Berm for critical reading of the manuscript and B. Phillips for technical assistance. We also thank H. Helm for computer graphics.

REFERENCES 1. 2. 3.

4.

5.

6.

8. 7.

9. 10.

11.

12.

13.

14. 15.

16.

17.

18.

19.

20.

21.

Mecsas, J., and Sugden, B. (1987) Annu. Reu. Cell Biol. 3,87-108 Ustav, M., and Stenlund, A. (1991) EMBO J. 10,449-457 Ustav. M.. Ustav. E.. Szvmanski. P.. and Stenlund. A. (1991) EMBO J.

10,’43zi-4329’ ’ “ , , . . .

Yang, L., Li, R., Mohr, I. J., Clark, R., and Botchan, M. R. (1991) Nature

Seo. Y. S., Mueller, F., Lusky, M., and Hurwitz, J. (1993) Proc. Natl. Acad.

Borowiec, J., Dean, F. B., Bullock, P., and Hurwitz, J. (1990) Cell 60,181-

353,628-632

Sci. U. S. A. 90, 702-706

1 RA Stiiiman, B. (1989) Annu. Reu. Cell Biol. 5 , 197-245 Challberg, M., and Kelly, T. (1989) Annu. Reu. Biochem 58,671-717 Winokur, P. L., and McBride, A. A. (1992) EMBO J. 11,4111-4118 Mohr, I. J., Clark, R., Shaw, S., Androphy, E. J., MacPherson, P., and

Lusky, M., and Fontane, E. (1991) Proc. Natl. Acad. Sei. U. S. A. 88,6363-

Seo, Y.-S., Muller, F., Lusky, M., Gibbs, E., Kim, H.-Y., Phillips, B., and

Kadonaga, J. T., and Tjian, R. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,

Botchan, M. R. (1990) Science 250,1694-1699

6367

Hurwitz, J. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 2865-2869

5889-5893 Wilson, V. G., and Ludes-Meyers, J. (1991) J . Virol. 65,5314-5322 Li, R., Knight, J., Bream, G., Stenlund, A., and Botchan, M. (1989) Genes

Hupp, T. R., Meek, D. W., Midgley, C. A,, and Lane, D. P. (1992) Cell 71, & Deu. 3,510-126

875-RR6 Sun S. Thorner L. Leutz, M., MacPherson, P., and Botchan, M. (1990)

Santucci, S., Androphy, E. Y., Bonne-Andrea, C., and Clertant, P. (1990)

Dean, F. B., Borowiec, J. A., Eki, T., and Hurwitz, J. (1992) J . Biol. Chem.

Mastrangelo, I . A., Hough, P. U. C., Wall, J. S., Dodson, M., Dean, F. B.,

Blitz, I. L., and Laimins, L. A. (1991) J. Virol. 65, 643-656

”_

J.’ d o l . 6 4 , 569315105

J. Virol. 64,6027-6033

2 6 7 , 14129-14137

and Hurwitz, J. (1989) Nature 338,658-662

cooperative assembly of the bovine papilloma virus e l and e2

Documents