Characterization of the Minimal DNA Binding Domain of the Human Papillomavirus E1 Helicase: Fluorescence Anisotropy Studies and Characterization of a Dimerization-Defective Mutant

JOURNAL OF VIROLOGY, May 2003, p. 5178–5191 Vol. 77, No. 90022-538X/03/$08.00�0 DOI: 10.1128/JVI.77.9.5178–5191.2003Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Characterization of the Minimal DNA Binding Domain of the HumanPapillomavirus E1 Helicase: Fluorescence Anisotropy Studies and

Characterization of a Dimerization-Defective Mutant ProteinS. Titolo, K. Brault, J. Majewski, P. W. White, and J. Archambault*

Department of Biological Sciences, Boehringer Ingelheim (Canada) Ltd., Laval, Canada H7S 2G5

Received 26 September 2002/Accepted 16 January 2003

The E1 helicase of papillomaviruses is required for replication of the viral double-stranded DNA genome, inconjunction with cellular factors. DNA replication is initiated at the viral origin by the assembly of E1monomers into oligomeric complexes that have unwinding activity. In vivo, this process is catalyzed by the viralE2 protein, which recruits E1 specifically at the origin. For bovine papillomavirus (BPV) E1 a minimalDNA-binding domain (DBD) has been identified N-terminal to the enzymatic domain. In this study, wecharacterized the DBD of human papillomavirus 11 (HPV11), HPV18, and BPV E1 using a quantitative DNAbinding assay based on fluorescence anisotropy. We found that the HPV11 DBD binds DNA with an affinity andsequence requirement comparable to those of the analogous domain of BPV but that the HPV18 DBD has ahigher affinity for nonspecific DNA. By comparing the DNA-binding properties of a dimerization-defectiveprotein to those of the wild type, we provide evidence that dimerization of the HPV11 DBD occurs only on twoappropriately positioned E1 binding-sites and contributes approximately a 10-fold increase in binding affinity.In contrast, the HPV11 E1 helicase purified as preformed hexamers binds DNA with little sequence specificity,similarly to a dimerization-defective DBD. Finally, we show that the amino acid substitution that preventsdimerization reduces the ability of a longer E1 protein to bind to the origin in vitro and to support transientHPV DNA replication in vivo, but has little effect on its ATPase activity or ability to oligomerize into hexamers.These results are discussed in light of a model of the assembly of replication-competent double hexameric E1complexes at the origin.

Papillomaviruses are a family of pathogenic viruses that in-duce benign and malignant hyperproliferative lesions of thedifferentiating squamous and mucosal epithelium (reviewed inreferences 13, 28, 51, and 68). Among the best-characterizedhuman papillomaviruses (HPV) are those types that infect theanogenital region and are associated with the development ofbenign warts (HPV type 6 [HPV6] and -11; low-risk types) orcancerous lesions (HPV16, -18, and -31; high-risk types).

The life cycle of papillomaviruses is tightly coupled to thecellular differentiation program that occurs in the epithelium(for a recent review, see reference 52). These viruses infect thebasal cell layer where they establish their small double-stranded DNA genome, 7.9 kbp in length, as a circular extra-chromosomal element in the nucleus of infected cells. Main-tenance of the viral genome in the infected cell is central to thelife cycle of papillomaviruses and their associated pathologies.Hence, interfering with this process has been considered avaluable strategy for the development of antiviral therapeuticagents.

Maintenance of the viral genome in infected cells requiresthe activity of E1 and E2, the two viral proteins necessary forreplication of the HPV genome in conjunction with the hostcell DNA replication machinery (reviewed in reference 12). E1is the replicative helicase of papillomavirus (reviewed in ref-erence 56) and shares extensive amino acid sequence and func-

tional homology with the related helicase, large T antigen, ofsimian virus 40 (SV40) and polyomavirus (14, 36). As an ini-tiator protein E1 acts both as a DNA binding protein to rec-ognize the viral origin of DNA replication and subsequently asa helicase to unwind the origin and the DNA ahead of thereplication fork (22, 31, 49). Binding of E1 to an 18-bp invert-ed-repeat element within the origin (27, 39, 53) is facilitated byits interaction with E2 (5, 6, 7, 21, 22, 33, 34, 40, 44, 46, 48, 50,65) a dimeric transcription-replication factor that binds withhigh affinity to sites located in the origin (reviewed in reference38). The preferred binding site for bovine papillomavirus(BPV) E1 in complex with E2 is the hexanucleotide sequence5�-ATTGTT-3�; four to six related sites are present in an over-lapping fashion in the origin of many papillomaviruses (9). TheE1-E2-DNA complex is comprised of an E2 dimer and twomolecules of E1 (11) and serves as a starting point for theassembly of larger oligomeric E1 complexes, either hexamersor double hexamers, that have unwinding activity (20, 47). Invitro, E1 binds to DNA with little sequence specificity in ab-sence of E2 but can nevertheless assemble into replication-competent oligomers at high protein concentrations. Hence,E2 serves both as a specificity factor to direct E1 to the originand as a loading factor to recruit additional E1 monomers andfavor their assembly. A role for heat shock proteins in stimu-lating the assembly of oligomeric E1 complexes has been re-ported (30).

Structure function studies on BPV and HPV E1 have beenuseful in dissecting the domain organization of this replicativehelicase. The C-terminal domain of HPV11 E1 (amino acids353 to 649; Fig. 1) was shown to be sufficient for oligomeriza-

* Corresponding author. Mailing address: Department of BiologicalSciences, Boehringer Ingelheim (Canada) Ltd., 2100 Cunard St., La-val, Canada H7S 2G5. Phone: (450) 682-4640. Fax: (450) 682-4642.E-mail: [email protected].

5178

on February 8, 2016 by guest

http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

tion into hexamers that have ATPase and unwinding activity(58, 63). This enzymatic domain of HPV E1 is also responsiblefor interaction with the host polymerase alpha primase and, ina mutually exclusive manner, with the transactivation domainof E2 (2, 15, 37, 59, 66, 67). An additional interaction betweenE1 and E2 has been reported that involves the DNA bindingdomains (DBDs) of both proteins (4, 10, 11, 29, 41). Thissecond interaction occurs only between the two proteins ofBPV, and not those of HPV, and appears to be relevant onlyin the context of the BPV origin where an E1 binding site isjuxtaposed close to an E2 binding site (5, 43). In this case theinteraction between the E1 and E2 DBDs serves to trigger theinteraction between the E2 transactivation domain and the E1enzymatic domain (23).

A minimal DBD in BPV E1 was identified in the N-terminalportion of the protein upstream of the enzymatic domain (11,29, 45, 57). This domain (amino acids 159 to 303), which bindsDNA as a dimer, was recently crystallized in native form aswell as bound to DNA (18, 19). These crystal structures re-vealed that the E1 DBD is related to the origin-binding do-main of SV40 large T antigen despite the fact that they sharelittle sequence similarity (18). The structure also revealed thattwo conserved hydrophilic regions shown by mutagenesis to beinvolved in DNA-binding (24, 62, 64) are folded into an ex-

tended loop and an �-helix, respectively, that together form apositively charged DNA-binding surface. The DNA-bindingloop interacts on one strand with the first three nucleotides,ATT, of the hexanucleotide binding site 5�-ATTGTT-3�, whilethe remaining three nucleotides on the opposite strand, AAC,are bound by the DNA-binding helix. The DNA-binding loopmakes all the base contacts including the major ones with thethymidine at position 2 of the E1 binding site (19). Somedistortion in the DNA occurs upon binding of a DBD dimerthat becomes more pronounced upon binding of a seconddimer and which ultimately may facilitate unwinding (19). Thecrystal structures also revealed a DBD dimer interface that wasvalidated by mutagenesis. Specifically, two amino acid substi-tutions, V202R and A206R, in �-helix 3, which forms part ofthis dimer interface, were shown to result in mutant proteinsthat bind DNA as monomers (19). The requirement and roleof this dimer interface during replication is currently unknown.It has been hypothesized that each monomer within the initialE1 dimer may nucleate the assembly of an hexameric helicase,resulting in the formation of a replication-competent doublehexamer bound at the origin (18).

For HPV11 E1 previous studies indicated that the regioncorresponding to the minimal DBD identified in BPV E1 wasrequired but not sufficient for binding to the origin (54, 58). In

FIG. 1. (A) Schematic diagram of the HPV11 E1 protein, 649 amino acids in length, showing the location of the helicase domain (amino acids353 to 649) and of the DNA-binding domain (DBD, amino acids 191 to 353). Within the helicase domain is shown the location of the three ATPasemotifs characteristics of superfamily 3 (SF3) of NTPases (25). Empty boxes represent the four regions, termed A-D, of similarity with the relatedhelicase Large T antigen of SV40 and polyomavirus (14). (B) Purified proteins used in this study. The purified DBDs, either as fusion proteins withGST and a hexahistidine-tag (GST-His-DBD), or as hexahistidine-tagged proteins (His-DBD), were separated by SDS-PAGE and stained withCoomassie blue. The papillomavirus type of each DBD is indicated above each lane. 11A251R is a mutant version of the HPV11 DBD carryingthe A251R amino acid substitution in the dimer interface. (C) DNA-binding activity of the purified GST-His-DBD fusion proteins in EMSA.EMSA were performed with the indicated concentrations of GST-His-DBD and with a radiolabeled duplex oligonucleotide probe either containing(E1BS) or lacking (Control) two E1BS. Protein-DNA complexes were analyzed on an 8% polyacrylamide gel and visualized by autoradiography.The positions of the free and bound probe (E1•DNA) are indicated.

VOL. 77, 2003 DNA BINDING DOMAIN OF E1 HELICASE 5179


http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

these studies, the C-terminal oligomerization/enzymatic do-main was additionally required, perhaps because oligomeriza-tion of the protein stabilizes its interaction with the origin.Studies using purified recombinant HPV11 E1 helicase sug-gested that this protein binds DNA with even less sequencespecificity than BPV E1 (17). However, this last study wasperformed with an E1 protein that was purified as preformedoligomers (17, 42) and which may bind to DNA differently thanmonomeric E1 or the isolated minimal DBD.

To determine if the region of HPV11 E1 corresponding tothe minimal DBD identified in BPV E1 is capable of bindingDNA in absence of the C-terminal enzymatic domain, we pro-duced this domain of HPV11 E1 and characterized its DNA-binding activity using a quantitative assay based on fluores-cence anisotropy. For comparison and to determine thegenerality of our findings, we similarly produced and charac-terized the minimal E1 DBD from BPV and from HPV18, ahigh-risk virus. We found that the HPV11 DBD binds DNAwith an affinity and sequence specificity comparable to those ofthe analogous domain of BPV, but that the HPV18 DBD hasa higher affinity for nonspecific DNA. For HPV11, we provideevidence that dimerization of the DBD increases its affinity forDNA and occurs only on binding to two appropriately posi-tioned E1 binding sites. In contrast, the HPV11 E1 helicasepurified as preformed hexamers binds to DNA with little se-quence specificity. Finally, we have found that a deleteriousamino acid substitution in the dimer interface, when intro-duced in the context of a functional E1 helicase, affects theability of the protein to bind to the origin and support transientHPV DNA replication. These results are discussed in light ofa model for the assembly of double hexamers at the origin.

MATERIALS AND METHODS

Expression plasmids. Plasmids to express the E1 DBD of HPV11, HPV18,and BPV E1, as a fusion with glutathione S-transferase (GST) and a six-histidinetag, were constructed by inserting BamHI-EcoRI-digested PCR fragments en-coding the various DBDs between the BamHI and EcoRI sites of plasmidpGEX-4T-1 (Amersham Pharmacia Biotech). Primers used for amplificationwere designed so as to encode EcoRI and BamHI restriction sites and to encodea six-histidine tag at the N terminus of the DBD. The following pairs of primerswere used: HPV11 E1 DBD (amino acids 191 to 353), 5�-GGCTGGATCCCATCACCATCACCATCACGACACATCAGGAATATTAGAATTACTAAAATG-3� and 5�-GGGGAATTCACTAGTCAGCCAAACTATGTTCAATAAC-3�; HPV18 E1 DBD (amino acids 197 to 359), 5�-GGCTGGATCCCATCACCATCACCATCACACCATAGCACAATTAAAAGACTTGTTAAAAGT-3�and 5�-GGGAATTCACTAATCATCTATTCCATGTTGTATAATAGTAAGTC-3�; BPV1 E1 DBD (amino acids 159 to 303), 5�-GGCTGGATCCCATCACCATCACCATCACGCTACAGTTTTTAAGCTGGGGCTCTTTAAATC-3� and5�-GGGGAATTCACTAGTTCAGAGTAGTTTGCGCCCGTATCCACTCAG-3�.

To express the A251R mutant HPV11 E1 DBD, the plasmid encoding thewild-type protein was mutagenized with the QuikChange site-directed mutagen-esis kit (Stratagene) according to the manufacturer’s instructions and with thefollowing pair of complementary oligonucleotides: 5�-CATAGCATAGCAGATCGATTTCAAAAGTTAATTG-3� and 5�-CAATTAACTTTTGAAATCGATCTGCTATGCTATG-3�.

Protein expression and purification. Expression vectors for the various GST-His-E1 DBD were introduced into Escherichia coli BL21 (Novagen). Bacterialcultures were grown in Luria-Bertani media at 37°C to an optical density of 0.5(at 595 nm) and then induced for 6 h by the addition of 1 mM IPTG (isopropyl-�-D-thiogalactopyranoside). Cells were then harvested and frozen at �80°C forlater purification.

For each protein, a thawed pellet from a 500-ml culture was resuspended in 5ml of lysis buffer (50 mM Tris [pH 7.6], 250 mM NaCl, 5 mM EDTA, 5 mMdithiothreitol [DTT], 10% glycerol, 0.1% NP-40, antipain [10 �g/ml], leupeptin

[2 �g/ml], pepstatin [1 �g/ml], aprotinin [2 �g/ml]) and sonicated three times for30 s each using a Tekmar Sonic Disruptor. The resulting cell lysate was clearedby centrifugation at 30,000 � g for 30 min and then incubated with 0.4 ml ofglutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) for 3 h at roomtemperature. Beads were washed with 50 bed volumes of HS buffer (50 mM Tris[pH 8.0], 1 M NaCl, 5 mM EDTA, 5 mM DTT, 10% glycerol) followed by 50 bedvolumes of LS buffer (same as HS buffer but containing 0.2 M NaCl). Fusionproteins were either eluted at room temperature for 30 min in 1 bed volume ofelution buffer (25 mM Tris [pH 8.0], 0.2 M NaCl, 1 mM EDTA, 5 mM DTT, 10%glycerol, 20 mM reduced glutathione [pH 8.0]) or cleaved with thrombin torelease the polyhistidine-tagged DBD from the GST-moiety. For thrombin cleav-age, bead-bound GST-His-DBD fusion proteins were resuspended in 1 bedvolume of cleavage buffer (25 mM Tris [pH 8.0], 0.2 M NaCl, 1 mM EDTA, 5mM DTT, 2.5 mM CaCl2, 10% glycerol) and incubated for 16 h at 4°C with 2.5units of biotinylated thrombin (Novagen). The supernatant was then recovered,and thrombin was removed by incubation with 100 �l of streptavidin agarosebeads (Novagen) for 30 min at 4°C. The soluble cleaved DBD was then recoveredafter centrifugation of the supernatant at 14,000 � g for 15 min at 4°C and storedat �80°C. Protein concentrations were determined by absorbance readings at280 nm in 6 M guanidine hydrochloride using the following calculated molarextinction coefficients of each of the following proteins: GST fusion proteins,HPV11 72.7 mM�1 cm�1, HPV18 70.9 mM�1 cm�1, and BPV1 60.1 mM�1

cm�1; purified DBDs, HPV11 31.5 mM�1 cm�1, HPV18 29.8 mM�1 cm�1, andBPV1 18.9 mM�1 cm�1.

Expression and purification of the truncated HPV11 E1(72-649) helicase,lacking the N-terminal 71 amino acids of the enzyme, was described previously(63).

Electrophoretic mobility shift assays (EMSA). Binding reactions were per-formed at room temperature in a total volume of 10 �l and in the followingbuffer: 20 mM Tris (pH 7.6), 100 mM NaCl, 1 mM EDTA, 0.1% NP-40, bovineserum albumin (1 mg/ml), 1 mM DTT, and 10% glycerol. Reactions contained 7nM radiolabeled probe and either 100, 50, 25 or 0 nM concentrations of GST-His-DBD protein. After 20 min samples were analyzed by electrophoresis on a0.5� Tris-borate-EDTA (TBE) polyacrylamide (8%) gel and visualized by au-toradiography.

To prepare the DNA probes, 2.5 �g of overlapping oligonucleotides wereannealed by heating at 70°C for 5 min in 100 �l of 1 � One-Phor-All Buffer Plus(Amersham Pharmacia Biotech) and then slowly cooled to room temperature.To facilitate labeling, oligonucleotides were designed to generate duplex DNAswith four protruding nucleotides at both 5� ends. For labeling, 1 �g of annealedoligonucleotides was radiolabeled with 30 U of Klenow fragment in a 120-�lreaction mixture using the same buffer as for annealing but supplemented with0.1 mM concentrations of dGTP, dATP, and dTTP and 100 �Ci of [�-33P]dCTPat 3,000 Ci/mmol. Labeling reactions were performed at room temperature for2 h, and probes were purified using MicroSpin G-25 columns (Amersham Phar-macia Biotech) according to the instructions supplied by the manufacturer.

One probe was designed such as to contain two inverted E1 binding sites(E1BS) separated by 3 bp. A similar probe in which the two E1BS were eachinactivated by two mutations was used as a control. The sequence of these twopairs of oligonucleotides was as follows (with the E1 binding sites underlined):2E1BS, 5�-CCGGGCGGGATTGTTGCTAACAATGGGCG-3� and 5�-CCGGCGCCCATTGTTAGCAACAATCCCGC-3�; control probe, 5�-CCGGGCGGGTTTCTTGCTAAGAAAGGGCG-3� and 5�-CCGGCGCCCTTTCTTAGCAAGAAACCCGC-3�.

Fluorescence anisotropy DNA binding assay. Binding assays were performedin 96-well HTRF plates (Packard) using 10 nM fluorescein-labeled probe and theindicated concentrations of protein in 150-�l reaction mixtures using the follow-ing buffer: 20 mM Tris (pH 7.6), 50 mM NaCl, 1 mM EDTA, 0.02% NP-40, and1 mM DTT. Fluorescence readings were taken on a POLARstar Galaxy platereader (BMG LabTechnologies GmbH) with excitation and emission filters at485 and 520 nm or on a Victor 2 1420 Multilabel HTS Counter (Wallac) usingthe 485 nm/535 nm filter sets. Background fluorescence from buffer was sub-tracted, and polarization and anisotropy values were defined as follows: P � (I�

� I[perp])/(I� � I[perp]) and A � (I� � I[perp])/(I� � 2I[perp]), where I� and I[perp] arethe fluorescence intensities recorded in the parallel and perpendicular orienta-tions respective to the orientation of the excitation polarizer.

Fluorescent DNA probes. High-pressure liquid chromatography-purified, flu-orescein-labeled oligonucleotides were purchased from Genset (Paris, France),with the fluorophore attached at the 5� end by a six-carbon linker. Blunt duplexDNA probes were prepared by annealing each fluorescein-labeled oligonucleo-tide to a complementary oligonucleotide. The sequence of the different fluores-cein-labeled oligonucleotides were as follows (the E1 binding sites are under-lined; “F” indicates the position of the fluorescein moiety):

5180 TITOLO ET AL. J. VIROL.


http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

1 bp 2E1BS, 5�-F-GATTGTTGCTAACAATGGGC-3�; 2 bp 2E1BS probe,5�-F-GGATTGTTGCTAACAATGGGC-3�; 4 bp 2E1BS probe, 5�-F-CGGGATTGTTGCTAACAATGGGC-3�; 2 bp proximal E1BS probe, 5�-F-GGATTGTTGCTAAGAAAGGGC-3�; 2 bp distal E1BS probe, 5�-F-GGTTTCTTGCTAACAATGGGC-3�; 0 E1BS probe, 5�-F-GGTTTCTTGCTAAGAAAGGGC-3�.

Competitor oligonucleotides. Unless stated otherwise, all duplex oligonucleo-tides that were used in competition experiments were designed similarly to thoseused in EMSA, so as to carry a 4-nucleotide overhangs at each 5� end. Oligo-nucleotides were designed to contain 1, 2, or no E1BS. Duplex oligonucleotidescontaining two and no E1BS, respectively, were the same as those used in EMSA(see above). The duplex DNA containing a single E1BS was made by annealingthe following two oligonucleotides (with the E1 binding sites underlined and theinactivating mutations written in bold): 5�-CCGGGCGGGTTTCTTGCTAACAATGGGCG-3� and 5�-CCGGCGCCCATTGTTAGCAAGAAACCCGC-3�. Incontrol experiments, we verified that the presence or absence of the 4 nucleotideoverhang had no effect on the affinity of the different DBDs for their targetsbinding site, yielding identical Ki values (data not shown).

To investigate the effect of the position of the E1BS relative to the end of theDNA duplex (see Fig. 4A), blunt duplex oligonucleotides were used in which thetwo E1BS were located 0 to 6 bp away from the duplex end. The sequence of thetop strand of these oligonucleotides was as follows (with the E1 binding sitesunderlined):

0 bp, 5�-ATTGTTGCTAACAATGGGCGGCGGGCCGG-3�; 1 bp, 5�-GATTGTTGCTAACAATGGGCGGCGGGCCG-3�; 2 bp, 5�-GGATTGTTGCTAACAATGGGCGGCGGGCC-3�; 3 bp, 5�-CGGATTGTTGCTAACAATGGCGGCGGGC-3�; 4 bp, 5�-CCGGATTGTTGCTAACAATGGGCGGCGGG-3�; 5bp, 5�-GCCGGATTGTTGCTAACAATGGGCGGCGG-3�; 6 bp, 5�-GGCCGGATTGTTGCTAACAATGGGCGGCG-3�.

Kd determination. Kd values were obtained from direct binding isotherms, witheach data point prepared as duplicates or quadruplicates, and fitted by nonlinearleast-squares regression with the program GraFit 3.09b (Erithicus Software Ltd.)to the standard equation describing the following tight binding equilibrium, D �E 7 DE (where D is duplex DNA, E is E1 DBD, and DE is DNA-DBDcomplex): A � Amax [(DT � ET � Kd) � {(DT � ET � Kd)2 � (4DTET)}1/2]/(2DT), where A is measured anisotropy gain at total concentrations of DBD(ET) and fluorescent probe (DT), Amax is anisotropy gain once probe is bound(Abound � Afree), and Kd is dissociation constant. No computational correctionsfor emission intensity were performed since the quantum yield did not changesignificantly upon binding of the DBD.

IC50 measurements and Ki determinations. The 50% inhibitory concentrations(IC50) of various DNA competitors was measured in the fluorescence polariza-tion assay. Competitors were titrated in binding reactions to reach 100% inhi-bition (each concentration point was performed in duplicate or quadruplicate, asindicated) in reactions performed at the indicated concentrations of protein andfluorescent probe. IC50 were determined by a nonlinear least-square regressionfit of the inhibition curve using the SAS program package (software release 6.12;SAS Institute Inc., Cary, N.C.). Ki values were then calculated by using thestandard equation that portrays the equilibrium EI7 I � E � D7 DE (whereI is inhibitor):

P0 � Pmax ([Kd(1 � IC50/Ki) � DT � ET] � { [Kd(1 � IC50/Ki) � DT � ET]2

� (4DTET)}1/2)/DT, where Pmax is the polarization gain once probe is bound

(Pbound � Pfree), P0 is the polarization gain in absence of inhibitor at theworking concentrations of DBD (ET) and probe (DT) and Ki the dissociationconstant of the inhibitor.

ATPase activity. ATPase activity of in vitro-translated E1 proteins was mea-sured as described previously (63). Briefly, radiolabeled E1 proteins that weresynthesized in vitro were immunoprecipitated using an anti-E1 antibody coupledto protein A beads. The ATPase activity of bead-bound immunopurified E1 wasthen assayed using an ammonium molybdate-malachite green colorimetric assay.ATPase activities above background levels were normalized to the amount of35S-labeled E1 protein present in each immunoprecipitate, as determined bySDS-PAGE analysis and autoradiography.

Cross-linking of in vitro-translated E1. Oligomerization of in vitro-translatedE1 was determined using a cross-linking assay as described previously (58).Briefly, 35S-labeled E1 protein was incubated in presence or absence of 50 ng/�lsingle-stranded DNA (60-mer, corresponding to nucleotides 7902 to 34 of theHPV11 origin) and then cross-linked with the sulfhydryl-reacting reagent bis-maleimidohexane (BMH) (Pierce). Cross-linking was performed by diluting thebinding reactions 13-fold with phosphate buffer (0.1 M pH 7.0) containing 100�M BMH and then stopped after 1 min by addition of DTT to a final concen-tration of 2.5 mM. Cross-linked E1 proteins were then immunoprecipitated withan anti E1 antibody and analyzed by gel electrophoresis on 3% Weber-Osbornpolyacrylamide gel (60) and autoradiography.

E1 DNA-binding coimmunoprecipitation assay. The DNA-binding activity ofin vitro-translated E1 was assayed as described previously (58). Briefly, in vitro-translated E1 protein was incubated with a 33P-radiolabeled DNA probe com-prised of two DNA fragments, one containing and the other lacking the minimalHPV11 origin of DNA replication. Binding reactions were allowed to proceed atthe indicated temperature for 90 min. When indicated, ATP and MgCl2 weresupplemented to the binding reactions at a final concentration of 5 and 3 mM,respectively. We previously noted that ATP-Mg is essential for binding whenreactions are performed at 37°C. DNA-protein complexes were immunoprecipi-tated using an anti-E1 antibody coupled to protein-A Sepharose beads, andwashed extensively. The coimmunoprecipitated DNA present in these complexeswas then analyzed on a 5% polyacrylamide TBE gel and visualized by autora-diography.

Transient HPV DNA replication assay. This assay was performed essentially asdescribed previously (58) but with the following modifications. Briefly, approxi-mately one million CHO-K1 cells were transfected using Lipofectamine (GibcoBRL) with three plasmids encoding, respectively, E1 (pCR3-E1; 500 ng), E2(pCR3-E2; 50 ng), and the minimal origin of DNA replication of HPV11 (pN9;500 ng). At 4 h posttransfection, cells were treated with trypsin and seeded in96-well plates (20 � 103 cells/well). Cells were harvested 48 h posttransfection,and total genomic DNA was isolated with the QIAmp blood kit (Qiagen).Replicated pN9 plasmid DNA was then detected by PCR amplification of anorigin-containing fragment using DpnI-digested total genomic DNA as templateand with the same primers as described previously (58). As a control, a fragmentof the pCR3-E1 plasmid devoid of Dpn1 restriction sites was amplified in thesame PCR using the following pair of primers that hybridize within the E1 openreading frame: 5�-ACCACATGTGCCGATTGGGTGGTTGCAGGA-3� andGCTGAAGGGTCACAGTCCACCGGGATGTT-3� (corresponding to nucleo-tides 1525 to 1554 and 2286 to 2258 of the HPV11 genome). The PCR conditionsconsisted of an initial denaturation step at 95°C for 5 min, followed by 21 roundsof sequential denaturation at 95°C for 30 s and extension at 72°C for 1.5 min, andending with a final extension at 72°C for 10 min. In control experiments, wefound that this low number of PCR cycles insures that amplification reactions forboth fragments remain in the linear range (data not shown). PCR products wereseparated on a 1% TBE agarose gel and visualized by staining with the interca-lating dye SYBRGreen I (Molecular Probes). The amount of replicated HPV(pN9) DNA was quantified by exposure on a Storm 860 Phosphorimager (Mo-lecular Dynamics) and normalized to the amplified E1 signal. Transfection anddetection of replicated pN9 plasmid were performed in quadruplicates.

RESULTS

Expression and preliminary characterization of the E1DNA-binding domain from HPV11 and HPV18. A minimalorigin DNA-binding domain (DBD) in BPV E1 has beenmapped between residues 159 and 303 (see the introductionfor references). Our previous analysis of HPV11 E1, usingprotein made by in vitro translation, indicated that a similarregion encompassing amino acids 191 to 353 (Fig. 1A) is nec-essary for binding to the viral origin (58). To determine ifresidues 191 to 353 of HPV11 E1 are sufficient for binding toDNA, we expressed and purified this domain as a fusion pro-tein with GST and a hexahistidine tag (Fig. 1B). We alsoexpressed and purified in a similar manner the correspondingdomain of HPV18 E1 (amino acids 197 to 359) and, as acontrol, that of BPV E1 (amino acids 159 to 303) (Fig. 1B).These purified GST-DBD proteins were then tested with anEMSA for binding to a duplex oligonucleotide probe contain-ing two inverted E1 binding sites (E1BS; 5�-ATTGTT-3�) sep-arated by 3 bp. This arrangement of E1BS was suggested re-cently to constitute the minimal binding sequence for a dimerof the BPV E1 DBD (9). As a control for specificity, theseproteins were also tested for binding to a mutant probe inwhich both E1BS were inactivated by two mutations each (seeMaterials and Methods). Throughout this study, we will referto the specific probe or related ones as containing two E1BS,and to the mutant probe as nonspecific DNA. As can be seen



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

in Fig. 1C, the GST-DBD fusion proteins, but not GST alone(data not shown), were able to bind the probe carrying twoE1BS, but not the control probe. These results indicate thatamino acids 191 to 353 of HPV11 E1 and the homologousdomain of HPV18 E1 (amino acids 197 to 359) encode asequence specific DNA-binding domain, similarly to what wasobserved previously for BPV E1.

We next set out to characterize the DNA-binding activity ofthe E1 DBD from HPV11, HPV18 and BPV in the absence ofGST as a fusion partner, since GST is itself a dimer. The GSTmoiety was removed from the different fusion proteins by pro-teolytic cleavage. The resulting hexahistidine-tagged DBDs(Fig. 1B) were found to be monomeric in solution under theconditions used in DNA-binding reactions, when analyzed bysize exclusion chromatography or analytical ultracentrifugation(data not shown). To better characterize the DNA-bindingactivity of these monomeric E1 DBDs, we then set out todevelop a quantitative assay based on fluorescence anisotropy,which unlike EMSA allows for measurements to be performedin solution and at equilibrium (26, 32). These studies are pre-sented below.

Measurement of the DNA binding activity of HPV11 E1 byfluorescence anisotropy. As a first step towards developing aDNA-binding assay based on fluorescence anisotropy, we mea-sured the binding of the HPV11 E1 DBD to four differentfluorescein-labeled duplex oligonucleotides. Three of theprobes contained two inverted E1BS separated by three bp,but positioned 1, 2, and 4 bp, respectively, away from thefluorescein-labeled end of the duplex (Fig. 2A). The fourthprobe was a control probe with both E1BS mutated. As can beseen in Fig. 2B, titration of the HPV11 E1 DBD resulted in adose-dependent change in anisotropy for all four probes. Thechange in anisotropy was maximal for the probe with E1BSpositioned 2 bp away from the fluorescein-labeled end of theduplex. From a series of binding isotherms similar to thatshown in Fig. 2B, a dissociation constant (Kd) of 10 nM wasobtained for the binding of the E1 DBD to this probe. Incontrast, the DBD bound to the probe with E1BS located 1 bpaway from the end of the duplex with a higher Kd of 50 nM andgave rise to a maximal change in anisotropy that was of inter-mediate value. Interestingly, binding of the E1 DBD to theprobe with the two E1BS positioned 4 bp away from the end ofthe duplex consistently generated a biphasic curve. It is con-ceivable that the lower gain in anisotropy (20 mP) seen at lowprotein concentrations is due to specific binding of the DBDon the two E1BS, whereas the larger anisotropy change is dueto nonspecific binding of additional DBD molecules to theprobe. As anticipated the E1 DBD bound only weakly to thecontrol probe indicating that its binding to DNA is sequence-specific. Collectively, these results indicate that the presence ofthe two E1BS and their distance relative to the fluorescein-labeled end of the duplex are critical parameters to detectbinding of the HPV11 E1 DBD to the probe. A spacing of 2 bpwas found to be optimal and hence was used in subsequentexperiments.

Evidence that the HPV11 E1 DBD binds DNA as a dimer. Toobtain evidence that the HPV11 E1 DBD binds DNA as adimer, we first measured its binding to probes in which eitherthe E1BS proximal or the one distal to the fluorescein moietywas mutated. In addition to the wild type DBD, we produced

and tested a mutant protein carrying a deleterious amino acidsubstitution, A251R, in the dimer interface (Fig. 1B). Theanalogous substitution in the BPV E1 DBD, A206R, wasshown recently to result in a protein that binds DNA as amonomer (18). Three observations were made that are consis-tent with the notion that the HPV11 E1 DBD binds DNA as adimer (Fig. 3). Firstly, binding of the wild type DBD to a probecarrying only the proximal E1BS (Kd � 150 nM) was muchweaker than to one with two E1BS (Kd � 10 nM). In addition,the fact that the change in anisotropy was lower for the probewith only the proximal E1BS than for one with two E1BS isconsistent with the binding of a single DBD molecule, ratherthan two, to this mutant probe. Secondly, the affinity of theA251R mutant protein for the probe with two E1BS was thesame as that for the probe with only the proximal E1BS, andcomparable to that of the wild type protein for the proximalE1BS. In this case also, the lower changes in anisotropy de-tected with the A251R protein are consistent with this proteinbinding to DNA as a monomer only. Thirdly, binding of theDBD to a probe carrying only the distal E1BS generated asubstantial change in anisotropy for the wild type DBD but notfor the mutant protein. We interpret this result to indicate thatonly binding events that occur proximal to the fluoresceinmoiety generate a large change in anisotropy, a suggestion thatis consistent with our initial observation that the spacing be-tween the E1BS and the fluorescein moiety is critical to detectspecific binding. In this scenario, binding of a monomer of theA251R DBD to the distal E1BS would give rise to a muchsmaller signal because it occurs too far from the fluorescein

FIG. 2. DNA-binding activity of the HPV11 E1 DBD detected byfluorescence anisotropy. (A) Schematic diagram of the duplex DNAprobe indicating the presence of the two E1 binding sites (E1BS) andof the fluorescein moiety (F). Different probes were designed in whichthe spacing between the E1BS and the fluorescein moiety was 1, 2, or4 bp. (B) Binding isotherms generated using 10 nM of each probe orwith a control probe lacking any E1BS (0 E1BS). Each binding iso-therm was performed in duplicate.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

moiety. In contrast, binding of the wild type DBD to this distalE1BS would generate a substantial anisotropy gain becausethis protein is able to dimerize and in doing so brings a secondDBD molecule in proximity to the fluorophore. Finally and asexpected, we found that both the wild type and A251R mutantDBD have a weak affinity for the control probe devoid ofE1BS. Collectively, these results support the notion thatdimerization of the HPV11 DBD increases its affinity for twoE1BS.

Binding affinity measurements for the HPV11 E1 DBD.Binding of the DBD to fluorescent probes can be used toestimate affinity constants but these experiments are compli-cated by the fact that the protein may bind more weakly to anE1BS positioned close to the fluorescein-labeled end of theduplex than to one flanked by additional sequences. In addi-tion, the proximity of the fluorophore may artifactually alterthe binding of the DBD for its target site. Therefore, we usedcompetition experiments with unmodified duplex oligonucleo-tides as a more rigorous method of measuring affinity.

i) Position of the E1BS. To determine the effect of theposition of the E1BS on affinity, we measured binding of theHPV11 DBD to duplex oligonucleotides carrying two E1BSsites positioned from 0 to 6 bp away from one end of the duplex(Fig. 4A). As anticipated the E1 DBD bound more weakly tosites located only 0 to 2 bp from the end of the duplex, than tosites located more distally. Hence the proximity of the E1BS tothe end of the duplex does weaken binding. Therefore, oligo-nucleotides in which the E1BS are positioned at least 4 bpfrom the end of the duplex were used for all subsequent com-petition experiments. Finally, the fact that the affinity of theprotein for sites located 2 bp away from the end of the duplexwas comparable to the Kd value measured in Fig. 2 with asimilar fluorescent probe, suggested that the fluorescein moi-ety does not have a substantial effect on binding.

ii) Affinity of the HPV11 wild type and A251R DBD. Next weused competition experiments to compare the affinities of theHPV11 DBD and A251R mutant derivative for DNA duplexescarrying two, one, or no E1BS (Fig. 4B). As expected the wildtype protein showed an approximately 10-fold-higher affinityfor two inverted E1BS, spaced by three nucleotides, than to asingle site. In contrast, but as expected for a dimerization-defective protein, the A251R DBD had a similar affinity forone or two E1BS, which was comparable to that of the wildtype protein for a single site. The affinity of both the wild typeand mutant DBD for a single E1BS was only about two- tothreefold higher than for nonspecific DNA.

iii) Affinity of the oligomeric E1 helicase. Next, we wished tocompare the DNA-binding affinity of the DBD to that of alonger E1 protein with helicase activity. For these experimentswe used a truncated form of E1 lacking the first 71 amino acids,E1(72-649), which we showed previously is hexameric in solu-tion and retains both helicase and ATPase activity (63). Titra-tion of purified E1(72-649) in presence of a fluorescent probecontaining two E1BS resulted in a substantial change in an-isotropy but the shape of the binding curve was inconsistentwith a single binding site (data not shown), perhaps because ofthe oligomeric nature of the protein. Although this complicatesderiving a true Kd value for this interaction and calculating Ki

values for competitor oligonucleotides, it did not prevent thedetermination of the relative affinity of E1(72-649) for duplexoligonucleotides carrying either 1, 2 or no E1BS. In competi-tion experiments, we found that E1(72-649) has a similar af-finity for one or two E1BS (IC50 � 12 to 14 nM) and a slightlyweaker affinity for nonspecific DNA (IC50 � 49 nM) (Fig. 4B).Hence, our hexameric E1 preparation binds DNA with littlesequence-specificity, in contrast to the DBD.

Identification of a consensus binding sequence for HPV11

FIG. 3. Binding of the wild type and A251R HPV11 DBDs to fluorescent DNA duplexes carrying two, one or no E1BS. Binding isotherms wereperformed with either the wild type or A251R HPV11 E1 DBD and with 10 nM of the indicated probe. The probe contained either two E1BS(white squares), only the E1BS proximal (white circles) or the one distal (black circles) to the fluorescein, or no E1BS (white triangles), asindicated. Kd of the wild type and A251R proteins were 6.6 0.7 and 113 13 nM, respectively, for the two E1BS probe and 94 15 and 97 18 nM, respectively, for the probe containing a single fluorescein-proximal E1BS.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

E1. To identify a consensus binding sequence for a dimer ofthe HPV11 E1 DBD, we performed a systematic mutationalanalysis of the hexanucleotide E1BS, 5�-ATTGTT-3�. In theseexperiments, we used as competitors a series of mutated DNAduplexes containing two E1BS (Fig. 5A). Each position of theE1BS was mutated to the other three possible nucleotides andeach mutation was introduced in both E1BS simultaneously.The three-nucleotide spacer region between the two E1BS wasalso mutagenized in a similar way. Results shown in Fig. 5Bindicated that positions 1, 2 and 4 of the E1BS could not bemutated without affecting binding. In contrast some mutationsat positions 3, 5 and 6 were tolerated. Specifically, a less thana twofold loss in binding affinity was observed when the Tresidue at position 3 was mutated to A or G or when the twoT residues at positions 5 and 6 were changed to a C. From

these results, the following consensus E1BS was obtained: 5�-AT(A/G/T)G(C/T)(C/T). As for the spacer region, we found itto be very tolerant to mutations. All three nucleotides could bechanged to any of the other three possible nucleotides with a�2-fold loss in affinity (Fig. 5B).

Next we investigated the effect of varying the length of thespacer between both E1BS. The results presented in Fig. 5B,showed that a duplex DNA containing two E1BS spaced by 3bp was the most potent competitor, indicating that this spacingis optimal for binding. Shortening the spacer region to 1 or 2bp, or lengthening it to 4 or 5, greatly reduced binding to theHPV11 E1 DBD. This importance of the spacer length lendsfurther support to the notion that dimerization of the DBD ontwo E1BS is critical for high-affinity binding.

The consensus binding sequence for a dimer of the E1 DBDdescribed above is present in two overlapping copies in theorigin of HPV11 and of other HPV types (9). In the HPV11origin these two copies are comprised, respectively, of E1BS 1and 3, and E1BS 2 and 4 (Fig. 6A). Competition experimentswere used to measure the affinity of the HPV11 E1 DBD for acombination of either E1BS 1 and 3 or E1BS 2 and 4 or for theorigin. As indicated in Fig. 6B, the binding sequence comprisedof E1BS 1 and 3 had the weakest affinity, comparable to thatfor a single consensus E1BS. In contrast, the combination ofE1BS 2 and 4 was nearly as good a competitor as the originsequences or two consensus E1BS. This was somewhat surpris-ing since E1BS 2 carries a G at position 4, which should bedetrimental to binding. However, it has been observed for BPVthat the presence of a deleterious G at position 4 can becompensated for by a thymidine at position 3 (19), as is thecase for E1BS 2. As suggested by Enemark et al. (19), suchcompensatory phenomenon is consistent with some mutationshaving more of a structural effect on DNA such as affecting itsability to become distorted.

Relative affinities of the E1 DBD from HPV11, HPV18 andBPV for duplex DNA. Competition experiments were used tocompare the affinities of the HPV11 DBD to those of HPV18and BPV, for duplex oligonucleotides carrying two, one or noE1BS.

(i) Affinity for 2, 1 or no E1BS. As indicated in Fig. 7, understandard assay conditions (i.e., 50 mM NaCl), the affinities ofthe three DBDs for two E1BS were comparable. In contrast,the HPV18 DBD showed a substantially higher affinity for aduplex DNA containing either one or no E1BS than the othertwo DBDs. In an attempt to reduce the higher nonspecificDNA-binding of the HPV18 DBD, these experiments wererepeated at a higher salt concentration (100 mM NaCl). Asanticipated, increasing the salt concentration reduced the af-finity of all three DBDs for any given duplex DNA. However,the HPV18 DBD still retained its higher affinity for nonspecificDNA compared to the other two DBDs.

(ii) Affinity of GST-fusion proteins. The finding that theHPV18 DBD binds to DNA with little sequence-specificity isin apparent contrast to what we observed in EMSA using aGST-fusion protein (Fig. 1). To determine the effect of theGST moiety, we measured in the fluorescence assay the affinityof the three GST-DBD fusion proteins for duplexes containing2, 1 or no E1BS. These experiments were performed undersimilar conditions as used for EMSA, in buffer containing 100mM NaCl. Under these conditions, the HPV18 DBD showed

FIG. 4. (A) Affinity of the HPV11 E1 DBD for two E1BS posi-tioned at various distances from the end of the DNA duplex. A dia-gram of the competitor duplex oligonucleotides used in this experi-ment is shown above the bar graph with a double arrow indicating theportion of the duplex that was varied in length from 0 to 6 bp. Ki valueswere determined by titrating the indicated competitor DNA duplexesin binding reactions containing 100 nM of HPV11 E1 DBD and 10 nMof a two E1BS fluorescent-probe (Kd for this probe was measured to be5 nM). Each Ki is the average of two independent measurements,which differed by less than 10%. (B) Affinity of the HPV11 E1 DBDand E1(72-649) helicase for duplex oligonucleotides carrying two, oneor no E1BS. Ki or IC50 values were determined by titrating duplexoligonucleotides containing 2, 1 or no E1BS in binding reactions con-taining 10 nM of a two E1BS fluorescent-probe. The concentration ofprotein in these binding reactions was 30 nM for the wild type DBD,450 nM for the A251R DBD, and 50 nM for E1(72-649). The Kd of thewild type and mutant DBD for the probe was measured to be 14 nMand 140 nM, respectively. Each Ki is the average of two independentmeasurements, which differed by less than 10%.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

a 15-fold preference for two E1BS (Kd � 7 nM) over nonspe-cific DNA (Kd � 105 nM), which accounts for its sequence-specificity observed in EMSA. However, even as a GST-fusionprotein, the HPV18 DBD still remained the least specific ofthe three DBDs, with the BPV and HPV11 fusion proteinsshowing a greater than 35-fold and 350-fold preference, re-spectively, for two E1BS over nonspecific DNA. Finally, wedetermined that the A251R GST-DBD bound to two E1BSwith lower affinity than its wild type counterpart, indicatingthat the interaction between DBDs is critical even in the con-text of a GST-fusion protein. Similar results were also ob-served in EMSA with this mutant protein (data not shown).

Effect of the A251R substitution on the replication functionsof HPV11 E1. The quantitative binding affinity measurementspresented above suggest that the A251R substitution affectsspecifically the dimerization of the HPV11 DBD but not theaffinity of a DBD monomer for DNA. To further characterizethe effect of this substitution on the replication activities of E1,we tested its effect on the ATPase, oligomerization and origin-binding activities of the protein in vitro. These experimentswere performed with in vitro-translated E1(72-649). Weshowed previously that this truncated protein, lacking the N-terminal 71 amino acids, is as active as wild type E1 in sup-porting cell-free HPV DNA replication but unlike the full-length protein binds more readily to the viral origin in vitro (1,58). We also tested the effect of the A251R substitution on the

ability of full-length E1 to support transient HPV DNA repli-cation in vivo. These results are presented below.

(i) ATPase activity. ATPase activity of in vitro-translated E1was measured using a colorimetric assay and following immu-noprecipitation of the protein from the translation reaction, asdescribed previously (63). In this assay, the A251R mutantprotein was found to be almost as active as the wild typeprotein (Fig. 8A). The previously characterized and catalyti-cally inactive K484R mutant E1 protein (63) was used as neg-ative control in this experiment.

(ii) Oligomerization. Oligomerization of in vitro-translatedand radiolabeled E1 was determined using a cross-linking as-say that we previously described (58). In this assay, singlestranded DNA is used to promote oligomerization of E1,which is then cross-linked with the sulfhydryl-reacting reagentBMH. After stopping the cross-linking reaction, the protein isimmunoprecipitated and analyzed by SDS-PAGE and autora-diography. As can be seen in Fig. 8B, the A251R mutant E1was as proficient as its wild type counterpart in assembling intohexamers. The previously characterized oligomerization defec-tive K484R and Y380A mutant E1 proteins (58) were used asnegative controls in this experiment.

(iii) Origin binding. To test for origin binding we used apreviously described DNA coimmunoprecipitation assay (58)in which in vitro-translated E1 is first incubated with two ra-diolabeled DNA fragments, one containing and the other lack-

FIG. 5. Effect of mutations in the E1BS on binding affinity. (A) Sequence of the two E1BS and spacer region into which mutations wereintroduced. Each mutation was introduced in both E1BS sites simultaneously. Each mutated duplex oligonucleotide was titrated as a competitorin binding reactions containing 30 nM of HPV11 E1 DBD and 10 nM of a two E1BS fluorescent-probe (Kd for this probe was measured to be 12nM). (B) Bar graph indicating the affinity of each mutated competitor DNA relative to that of the wild type (WT) DNA duplex. Each position ofthe two E1BS was changed to the other three nucleotides, indicated below each bar. For experiments in which the length of the spacer region wasvaried, the sequence of the altered spacer region is indicated below the bar. IC50 values of competitor DNAs were determined in quadruplicateand are presented, along with standard deviations, relative to that of the unaltered wild type (WT) sequence set arbitrarily at 1.0.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

ing the origin. E1-DNA complexes are then immunoprecipi-tated with an anti-E1 antibody and the coprecipitated DNAanalyzed by gel electrophoresis and autoradiography. As canbe seen in Fig. 8C, the A251R mutant protein showed a re-duced ability to bind to the origin compared to wild type E1, aresult suggesting that the DBD-dimer interface is required forstable binding to the origin in this assay. The previously char-

acterized K484R and Y380A mutant E1 proteins (58) wereused as negative controls in this experiment. As observed pre-viously, we found that the K484R mutant protein retains someactivity at 23° that is lost at 37°C (58).

(iv) Transient HPV DNA replication. A cotransfection assaywas used to measure the ability of the A251R mutant E1 tosupport transient HPV DNA replication. In this assay, con-

FIG. 6. Affinity of the HPV11 E1 DBD for its cognate origin. (A) Sequence of a portion of the HPV11 origin showing the location of E1BS1 to 4 and of the AT-rich region. (B) Affinity of the HPV11 E1 DBD for the origin or for subsets of E1BS. Ki values were determined by titratingduplex oligonucleotides spanning the origin, or containing the indicated combination of E1BS, in binding reactions containing 30 nM of HPV11E1 DBD and 10 nM of a two E1BS fluorescent-probe (Kd for this probe was measured to be 14 nM). The sequence of one strand of each duplexoligonucleotide is indicated. Each Ki is the average of two independent measurements, which differed by less than 10%.

FIG. 7. Affinity of the HPV11, HPV18 and BPV E1 DBD for duplex oligonucleotides carrying two, one or no E1BS. Ki values were determinedby titrating duplex oligonucleotides containing 2, 1, or no E1BS in binding reactions containing 10 nM of a two E1BS fluorescent probe, in buffercontaining either 50 mM or 100 mM NaCl, as indicated. At 50 mM NaCl, a concentration of 15 nM of each His-DBD was used. Under theseconditions, the measured Kd of the HPV11, HPV18 and BPV DBD for the two E1BS probe were of 5 nM, 2 nM and 6 nM, respectively. At 100mM NaCl, the HPV11, HPV18 and BPV His-DBD were used at a concentration of 125 nM, 50 nM and 200 nM, respectively. Under theseconditions, the measured Kd of the HPV11, HPV18 and BPV His-DBD for the two E1BS probe were of 131 nM, 18 nM and 82 nM, respectively.Ki values for the GST-His-DBD fusion proteins were measured using 20 nM of each protein in buffer containing 100 mM NaCl. Under theseconditions, the Kd of the HPV11, HPV18 and BPV GST-His-DBD for the two E1BS probes were measured to be 3 nM, 5 nM, and 13 nM,respectively. Each Ki is the average of two independent measurements, which differed by less than 10%.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

structs expressing E1 and E2 were transfected into CHO cellsalong with a plasmid carrying the HPV11 minimal origin ofDNA replication. 48 h post transfection, the amount of repli-cated DNA (Dpn1 resistant) was quantified by PCR. In thisassay, the A251R mutation greatly reduced the ability of E1 tosupport HPV DNA replication (Fig. 8D), consistent with theresults presented above that this mutation has a deleteriouseffect on origin-binding in vitro. The two DNA-binding defec-tive E1 proteins (K286A/R288A and A292L/R293E [58]) wereused as negative controls in this experiment.

DISCUSSION

In this study, we have characterized the E1 DBD of HPV11,HPV18 and BPV using a novel quantitative DNA-binding as-say based on fluorescence anisotropy. A key observation in thedevelopment of this assay was the realization that the positionof the E1BS relative to the fluorescein-labeled end of theduplex was critical to detect sequence-specific binding. Thisand other experiments suggested that only binding events thatoccur close to the fluorophore generate a large change in

FIG. 8. Effect of the A251R amino acid substitution on the replicative activities of HPV11 E1. (A) ATPase assay. ATPase activity of invitro-translated E1 (amino acids 72 to 649), either wild type (WT) or the indicated mutant derivative was measured following immunoprecipitationof the protein from the translation reaction. ATPase activity was normalized to the amount of immunoprecipitated protein and is presented relativeto the activity of the wild type protein set arbitrarily at 1.0. (B) Oligomerization assay. Oligomerization of radiolabeled E1 (amino acids 72 to 649)either wild type (WT) or the indicated mutant derivative was measured using a protein cross-linking assay. Proteins were cross-linked with BMHin presence of single stranded DNA, which stimulates oligomerization of E1. Proteins were then immunoprecipitated and analyzed by electro-phoresis on an a Weber-Osborn gel followed by autoradiography. The positions of monomeric (Mono) and hexameric (Hexa) E1, as determinedpreviously (58), are indicated. (C) Origin-binding assay. In vitro-translated E1 (amino acids 72 to 649), either wild type (WT) or the indicatedmutant derivative, was incubated with a radiolabeled DNA probe comprised of an origin-containing fragment (Ori, indicated by an arrow) and acontrol fragment. E1-DNA complexes were immunoprecipitated with an anti-E1 polyclonal antibody and the coprecipitated DNA analyzed by gelelectrophoresis and autoradiography. Binding reactions were performed at the indicated temperature in presence (�) or absence (-) of supple-mented ATP (5 mM). (D) Transient HPV11 DNA replication assay. Cells were transected with expression plasmids for E2 and the indicated wildtype or mutant full-length E1 protein, as well as an HPV11 origin-containing plasmid. Replication of the origin-containing plasmid (ori signalindicated by an arrow) was quantified by PCR on Dpn1-digested genomic DNA, and normalized to the amount of E1-expression plasmid presentin the transfected cells (E1 signal indicated by an arrow). The amount of HPV replication detected in cells transfected with wild type E1 and E2was set at 100%. As a negative control, replication of the origin-containing plasmid was measured in cells transected with E1 alone, in absence ofE2.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

anisotropy. As opposed to EMSA, this assay is performedunder true equilibrium conditions and allows for the measure-ments of accurate affinity values by competition experiments.The study presented here was initiated to provide a quantita-tive framework within which to understand the binding of E1from different papillomavirus types to their cognate origin, anessential step in the initiation of viral DNA replication. In thisrespect, our work complements previous fluorescence anisot-ropy studies that have measured the affinities of HPV11 E2and the E1-E2 complex for their cognate binding sites as wellas determined the stoichiometry of the E1-E2 complex (1, 8).Our study was also performed to help reconcile apparent dif-ferences between the E1 proteins of HPV11 and BPV. Specif-ically it has been reported that the binding of the HPV11protein to DNA is less sequence-specific than that of the BPVenzyme and requires the C-terminal enzymatic domain of theprotein in addition to the DBD (17, 53, 58). Results presentedhere clearly indicate that the HPV11 and BPV E1 proteins arenot dissimilar in that their respective DBD bind to DNA withvery similar affinities and sequence specificity. We now believethat the need for the C-terminal domain that we observedpreviously (58) was due to the fact that these studies wereperformed with in vitro-translated protein, which can only beobtained in low concentrations. Under these conditions, andgiven the relatively low affinity and sequence-specificity of mo-nomeric E1 for DNA (this study), the C-terminal domain maybe required to stabilize the association of in vitro-translated E1with the origin, perhaps by promoting its oligomerizationand/or by providing additional DNA contacts. As for the ap-parent lower sequence specificity reported earlier for HPV11E1 (17), it was probably due to the fact that the protein used inthis study was purified as preformed oligomers. Indeed wehave shown here that our preparation of enzymatically activeHPV11 E1, which is also purified as preformed hexamers (63),has little sequence-specificity showing only a fourfold-higheraffinity for one or two E1BS than for nonspecific DNA (Fig.4B).

In contrast to hexameric E1, the HPV11 DBD shows agreater than 20-fold higher affinity for two appropriately posi-tioned E1BS (Ki � 4 nM) than for nonspecific DNA (Ki � 96nM). The affinity of the DBD for a single E1BS (Ki � 48 nM)however is only two- to threefold higher than for nonspecificDNA, and in that respect the DBD behaves similarly to hex-americ E1. These results suggested that dimerization of theDBD is critical to achieve both sequence-specific and high-affinity binding. Characterization of a mutant HPV11 DBDcarrying a deleterious substitution, A251R, in the dimer inter-face provided additional evidence for the role of dimerization.In particular the following observations suggested that dimer-ization occurs only on binding of the DBD to two appropriatelypositioned E1BS. First, in solution, the DBD is monomeric atconcentrations of up to 4 mg/ml (data not shown), the highestconcentration tested, indicating that it does not dimerizereadily in absence of DNA. Second, the binding of the DBD toa fluorescent probe containing a single E1BS generated a max-imal change in anisotropy that was smaller than that observedwith a probe containing two E1BS, but similar to what isobserved for the binding of the dimerization-defective A251RDBD to either probes. The simplest interpretation of theseresults is that the lower gain in anisotropy results from the

binding of a single DBD to the probe whereas the larger oneresults from dimerization of the DBD on two E1BS. Third, theaffinity of the wild type DBD for a single E1BS (Ki � 45 nM)is comparable to that of the A251R mutant (Ki � 59 nM) andboth proteins have comparable affinities for nonspecific DNA(KI � 113 and 135 nM, respectively). If the wild type DBDcould dimerize readily in absence of E1BS, we would haveexpected its affinity for nonspecific DNA to be significantlyhigher than that of the A251R mutant protein. Fourth, thelength of the spacer region between two E1BS is critical for theaffinity of the wild type DBD (Fig. 5) but has little effect on thatof the A251R mutant protein. Altogether, the results pre-sented above indicate that dimerization of the DBD increasesboth sequence-specificity and affinity for its target site. Inagreement with this, we found that the HPV11 DBD has asubstantially higher affinity for two E1BS when artificiallymade into a dimer by fusion with GST, than as a monomericprotein (Ki of 2 and 80 nM, respectively; compare Fig. 7).

Our comparison of the HPV11 and BPV DBDs revealedthat their affinities for DNA are very similar (Fig. 7), with bothshowing a 10- to 20-fold higher affinity for two appropriatelypositioned E1BS than for nonspecific DNA. Both proteins alsoshow very little preference for a single E1BS compare to non-specific DNA. This is particularly striking for the BPV DBDwhich shows a �2-fold difference in affinity for duplex DNAscontaining one or no E1BS (Fig. 7), even as a dimeric GST-fusion protein (Fig. 7). Again, these results reinforce the no-tion that dimerization of the DBD on two E1BS is critical forformation of a stable DBD-DNA complex. From a biologicalstandpoint, the fact that formation of a stable dimer occursonly on two E1BS and that the DBD monomer has little affinityfor a single E1BS would help ensure that monomeric E1 doesnot get trapped on binding sites located outside of the origin.

The similarity between the HPV11 and BPV E1 DBD alsoextends to their sequence requirement. Indeed, our mutationalanalysis of the E1BS indicated that the sequence-specificity ofthe HPV11 E1 DBD is nearly identical to that previouslypublished for the BPV E1 DBD (9). This is despite the factthat the study for BPV was performed in the presence of theE2 DBD, although earlier studies have suggested that thesequence recognition of BPV E1 is not affected by its interac-tion with E2. As mentioned in the introduction, the interactionbetween the respective DBD of E1 and E2 does not appear toexist for HPV11 and, accordingly, we have been unable todetect an interaction between the HPV11 E1 DBD and E2 byEMSA on an origin-containing DNA probe (data not shown).To further characterize the sequence-specificity of the HPV11DBD, we examined the effect of mutations within the spacerregion between two inverted E1BS. These studies revealed thata spacing of any 3 bp was required for high affinity binding.

In contrast to the BPV and HPV11 E1 DBDs, we found thatthe analogous domain of HPV18 has a higher affinity for non-specific DNA and shows less of a sequence-preference, even asa GST-fusion protein. This suggests that the HPV18 DBDeither can dimerize more readily on DNA or that each mono-mer makes stronger contacts with the DNA backbone, or both.Further studies will be required to distinguish between thesepossibilities. Nevertheless, these findings reveal that there canbe subtle differences in how the E1 proteins from differentpapillomavirus types interact with their cognate origin. It was



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

previously shown for HPV18 that two E2 binding sites aloneare sufficient to function as a minimal origin of replication (55).It is likely that the lack of sequence specificity of HPV18 E1contributes to promoting DNA replication in absence of con-sensus E1BS for this HPV type.

In this study, we examined for the first time the effect of adeleterious amino acid substitution in the DBD dimer inter-face on the replicative activities of E1. We found that theA251R substitution does not affect the ATPase activity of E1nor its ability to oligomerize in presence of single strandedDNA. However, this substitution did reduce binding of E1 tothe origin in vitro and its ability to support transient HPVDNA replication in vivo, but did not abolish completely eitherof these activities. Hence these results suggest that the inter-action between DBDs is important but not essential for E1 tobind to the origin in vitro or for DNA replication in vivo.Enemark et al. (18) hypothesized that each monomer withinthe initial E1 dimer could nucleate the assembly of an hexa-meric helicase. Our findings that the A251R substitution af-fects neither the ability of the protein to assemble into hexam-ers in a cross-linking assay nor the interaction of themonomeric DBD with DNA per se, but specifically its ability todimerize, are entirely consistent with this proposal. Hence itmight be instructive to consider our results in light of a modelof the assembly of double-hexameric E1 complexes that in-volves the interaction between DBDs (Fig. 9). Based on ourfindings that the DBD dimerizes only upon binding to twoappropriately positioned E1BS, we propose that a major func-tion of the E1BS within the viral origin is to favor dimerizationof E1, through the DBD, such as to stabilize its initial binding

to the origin. Key to this process is the fact that monomeric E1is recruited to the origin by E2, such that the local E1 concen-tration at the origin is sufficiently high to favor dimerization.For HPV11, a dimer of E1 would probably assemble first onthe high affinity E1BS 2 and 4 followed by the assembly of asecond one on the lower affinity E1BS 1 and 3. Formation ofthese two E1 dimers would then act as a starting point for theassembly of double-hexamers, in a process that is likely torequire additional E1-E1 interactions, such as those involvedin oligomerization of the C-terminal enzymatic domain, and bestimulated by ATP. For the A251R mutant protein, interactionwithin the initial dimer cannot occur and as a result, binding ofthe first four E1 molecules to the origin would neither becooperative nor involve all four E1BS. This mutant proteincould still assemble into hexamers as suggested by this study,but these would be prevented from interacting with each otherto form a double hexamer. If this model were correct, the factthat the A251R mutant protein has residual activity wouldindicate that the interaction between E1 monomers leading tothe assembly of double hexamers is not essential for DNAreplication but contributes to its overall efficiency. The tran-sient HPV DNA replication assay used in these experimentsprobably measures no more than two rounds of DNA synthe-sis. It is conceivable that the interaction between DBDs may bemore critical for the long-term maintenance of the viral ge-nome in infected cells, which requires multiple rounds of DNAreplication.

For SV40 large T antigen, it has been suggested that theDBD plays a role in double hexamer assembly and that thisprocess is favored by phosphorylation of Thr124 by a cylin-cdkcomplex (3, 61). By analogy, one might speculate that thereported phosphorylation of E1 by cyclin E-cdk2 (16, 35) couldalso serve to promote double hexamer formation. It may there-fore not be a coincidence that substitutions in the E1 cyclin-binding-motif, like the A251R substitution, reduce but do notabolish transient HPV DNA replication in vivo (35; K. Braultand J. Archambault, unpublished observation)

ACKNOWLEDGMENTS

We thank Craig Fenwick and Steve Mason for critical reading of themanuscript.

REFERENCES

1. Alexander, K. A., and W. C. Phelps. 1996. A fluorescence anisotropy study ofDNA binding by HPV-11 E2C: a hierarchy of E2-binding sites. Biochemistry35:9864–9872.

2. Amin, A. A., S. Titolo, A. Pelletier, D. Fink, M. G. Cordingley, and J.Archambault. 2000. Identification of domains of the HPV11 E1 proteinrequired for DNA replication in vitro. Virology 272:137–150.

3. Barbaro, B. A., K. R. Sreekumar, D. R. Winters, A. E. Prack, and P. A.Bullock. 2000. Phosphorylation of simian virus 40 T antigen on Thr 124selectively promotes double-hexamer formation on subfragments of the viralcore origin. J. Virol. 74:8601–8613.

4. Benson, J. D., and P. M. Howley. 1995. Amino-terminal domains of thebovine papillomavirus type 1 E1 and E2 proteins participate in complexformation. J. Virol. 69:4364–4372.

5. Berg, M., and A. Stenlund. 1997. Functional interactions between papillo-mavirus E1 and E2 proteins. J. Virol. 71:3853–3863.

6. Blitz, I. L., and L. A. Laimins. 1991. The 68-kilodalton E1 protein of bovinepapillomavirus is a DNA binding phosphoprotein which associates with theE2 transcriptional activator in vitro. J. Virol. 65:649–656.

7. Bonne-Andrea, C., F. Tillier, G. D. McShan, V. G. Wilson, and P. Clertant.1997. Bovine papillomavirus type 1 DNA replication: the transcriptionalactivator E2 acts in vitro as a specificity factor. J. Virol. 71:6805–6815.

8. Chao, S. F., W. J. Rocque, S. Daniel, L. E. Czyzyk, W. C. Phelps, and K. A.Alexander. 1999. Subunit affinities and stoichiometries of the human papil-lomavirus type 11 E1:E2:DNA complex. Biochemistry 38:4586–4594.

FIG. 9. Model of the initiation of HPV DNA replication that em-phasizes the role of the DBD-dimer interface in assembly of oligo-meric E1 complexes at the origin. E1 is diagrammed as a two-domainprotein with the DBD shown as a gray circle and the C-terminalenzymatic (ATPase) domain indicated as a white ellipse. The modelhighlights the assembly of the two initial E1 dimers stabilized throughinteraction of their DBDs, on two pairs of overlapping E1BS, leadingto the assembly of replication-competent double-hexamers. For theA251R mutant E1, the presence of the arginine (R) residue thatprevents dimerization is indicated. See text for more details about thismodel.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

9. Chen, G., and A. Stenlund. 2001. The E1 initiator recognizes multiple over-lapping sites in the papillomavirus origin of DNA replication. J. Virol.75:292–302.

10. Chen, G., and A. Stenlund. 2000. Two patches of amino acids on the E2DNA binding domain define the surface for interaction with E1. J. Virol.74:1506–1512.

11. Chen, G., and A. Stenlund. 1998. Characterization of the DNA-bindingdomain of the bovine papillomavirus replication initiator E1. J. Virol. 72:2567–2576.

12. Chow, L. T., and T. R. Broker. 1994. Papillomavirus DNA replication. In-tervirology 37:150–158.

13. Chow, L. T., and T. R. Broker. 1997. Small DNA tumor viruses, p. 267–301.In N. Nathanson (ed.), Viral pathogenesis. Lippincott-Raven, Philadelphia,Pa.

14. Clertant, P., and I. Seif. 1984. A common function for polyoma virus large-Tand papillomavirus E1 proteins? Nature 311:276–279.

15. Conger, K. L., J.-S. Liu, S.-R. Kuo, L. T. Chow, and T. S.-F. Wang. 1999.Human Papillomavirus DNA Replication: Interactions between the viral E1protein and two subunits of human polymerase � primase. J. Biol. Chem.274:2696–2705.

16. Cueille, N., R. Nougarede, F. Mechali, M. Philippe, and C. Bonne-Andrea.1998. Functional interaction between the bovine papillomavirus virus type 1replicative helicase E1 and cyclin E-Cdk2. J. Virol. 72:7255–7262.

17. Dixon, E. P., G. L. Pahel, W. J. Rocque, J. A. Barnes, D. C. Lobe, M. H.Hanlon, K. A. Alexander, S. F. Chao, K. Lindley, and W. C. Phelps. 2000.The E1 helicase of human papillomavirus type 11 binds to the origin ofreplication with low sequence specificity. Virology 270:345–357.

18. Enemark, E. J., G. Chen, D. E. Vaughn, A. Stenlund, and L. Joshua-Tor.2000. Crystal structure of the DNA binding domain of the replication initi-ation protein E1 from papillomavirus. Mol. Cell 6:149–158.

19. Enemark, E. J., A. Stenlund, and L. Joshua-Tor. 2002. Crystal structures oftwo intermediates in the assembly of the papillomavirus replication initiationcomplex. EMBO J. 21:1487–1496.

20. Fouts, E. T., E. H. Egelman, and M. R. Botchan. 1999. Biochemical andelectron microscopic image analysis of the hexameric E1 helicase. J. Biol.Chem. 274:4447–4458.

21. Frattini, M. G., and L. A. Laimins. 1994. Binding of the human papilloma-virus E1 origin-recognition protein is regulated through complex formationwith the E2 enhancer-binding protein. Proc. Natl. Acad. Sci. USA 91:12398–12402.

22. Gillette, T. G., M. Lusky, and J. A. Borowiec. 1994. Induction of structuralchanges in the bovine papillomavirus type 1 origin of replication by the viralE1 and E2 proteins. Proc. Natl. Acad. Sci. USA 91:8846–8850.

23. Gillitzer, E., G. Chen, G., and A. Stenlund. 2000. Separate domains in E1and E2 proteins serve architectural and productive roles for cooperativeDNA binding. EMBO J. 19:3069–3079.

24. Gonzalez, A., C. Bazaldua-Hernandez, M. West, K. Woytek, and V. G. Wil-son. 2000. Identification of a short, hydrophilic amino acid sequence criticalfor origin recognition by the bovine papillomavirus E1 protein. J. Virol.74:245–253.

25. Gorbalenya, A. E., E. V. Koonin, and Y. I. Wolf. 1990. A new superfamily ofputative NTP-binding domains encoded by genomes of small DNA andRNA viruses. FEBS Lett. 262:145–148.

26. Heyduk, T., Y. Ma, H. Tang, and R. H. Ebright. 1996. Fluorescence aniso-tropy: rapid, quantitative assay for protein-DNA and protein-protein inter-action. Methods Enzymol. 274:492–503.

27. Holt, S. E., and V. G. Wilson. 1995. Mutational analysis of the 18-base-pairinverted repeat element at the bovine papillomavirus origin of replication:identification of critical sequences for E1 binding and in vivo replication.J. Virol. 69:6525–6532.

28. Howley, P. M. 1996. Papillomavirinae: the viruses and their replication, p.2045–2076. In B. N. Fields and D. M. Knipe (ed.), Fields virology, 3rd ed.Raven Press, New York, N.Y.

29. Leng, X., J. H. Ludes-Meyer, and V. G. Wilson. 1997. Isolation of an amino-terminal region of bovine papillomavirus type 1 E1 protein that retains originbinding and E2 interaction capacity. J. Virol. 71:848–852.

30. Liu, J.-S, S.-R. Kuo, A. M. Makov, D. M. Cyr, J. D. Griffith, T. R. Broker, andL. T. Chow. 1998. Human Hsp70 and Hsp40 chaperone proteins facilitatehuman papillomavirus-11 E1 protein binding to the origin and stimulatecell-free DNA replication. J. Biol. Chem. 273:30704–30712.

31. Liu, J.-S., S.-R. Kuo, T. R. Broker, and L. T. Chow. 1995. The functions ofhuman papillomavirus type 11 E1, E2, and E2C proteins in cell-free DNAreplication. J. Biol. Chem. 270:27283–27291.

32. Lundbald, J. R., M. Laurence, and R. H. Goodman. 1996. Fluorescencepolarization analysis of protein-DNA and protein-protein interactions. Mol.Endocrinol. 10:607–612.

33. Lusky, M., J. Hurwitz, and Y.-S. Seo. 1993. Cooperative assembly of thebovine papilloma virus E1 and E2 proteins on the replication origin requiresan intact E2 binding site. J. Biol. Chem. 268:15795–15803.

34. Lusky, M., J. Hurwitz, and Y.-S. Seo. 1994. The bovine papillomavirus E2protein modulates the assembly of but is not stably maintained in a replica-

tion-competent multimeric E1-replication origin complex. Proc. Natl. Acad.Sci. USA 91:8895–8899.

35. Ma, T., N. Zou, B. Y. Lin, L. T. Chow, and J. W. Harper. 1999. Interactionbetween cyclin-dependent kinases and human papillomavirus replication-initiation protein E1 is required for efficient viral replication. Proc. Natl.Acad. Sci. USA 96:382–387.

36. Mansky, K. C., A. Batiza, and P. F. Lambert. 1997. Bovine papillomavirustype 1 E1 and simian virus 40 large T-antigen share regions of sequencesimilarity required for multiple functions. J. Virol. 71:7600–7608.

37. Masterson, P. J., M. A. Stanley, A. P. Lewis, and M. A. Romanos. 1998. AC-terminal helicase domain of the human papillomavirus E1 protein bindsE2 and the DNA polymerase alpha-primase p68 subunit. J. Virol. 72:7407–7419.

38. McBride, A., and G. Myers. 1997. The E2 proteins, p. 54–73. In G. Myers, C.Baker, K. Munger, F. Sverdrup, A. McBride, and H.-U. Bernard (ed.),Human papillomaviruses 1997. Theoretical Biology and Biophysics, LosAlamos National Laboratory, Los Alamos, N.Mex.

39. Mendoza, R., L. Gandhi, and M. R. Botchan. 1995. E1 recognition sequencesin the bovine papillomavirus type 1 origin of DNA replication: interactionbetween half sites of the inverted repeats. J. Virol. 69:3789–3798.

40. Mohr, I. J., R. Clark, S. Sun, E. J. Androphy, P. MacPherson, and M. R.Botchan. 1990. Targeting the E1 replication protein to the papillomavirusorigin of replication by complex formation with the E2 transactivator. Sci-ence 250:1694–1699.

41. Moscufo., N., F. Sverdrup, D. E. Breiding, and E. J. Androphy. 1999. Twodistincts regions of the BPV1 E1 replication protein interact with the acti-vation domain of E2. Virus Res. 65:141–154.

42. Rocque, W. J., D. J. T. Porter, J. A. Barnes, E. P. Dixon, D. C. Lobe, J. L. Su,D. H. Willard, R. Gaillard, J. P. Condreay, W. C. Clay, C. R. Hoffman, L. K.Overton, G. Pahel, T. A. Kost, and W. C. Phelps. 2000. Replication-associ-ated activities of purified human papillomavirus type 11 E1 helicase. ProteinExpress. Purif. 18:148–159.

43. Sanders, C. M., and A. Stenlund. 2001. Mechanism and requirements forbovine papillomavirus, type 1, E1 initiator complex assembly promoted bythe E2 transcription factor bound to distal sites. J. Biol. Chem. 276:23689–23699.

44. Sanders, C. M., and A. Stenlund. 1998. Recruitment and loading of the E1initiator protein: an ATP-dependent process catalysed by a transcriptionfactor. EMBO J. 17:7044–7055.

45. Sarafi, T. R., and A. A. McBride. 1995. Domains of the BPV-1 E1 replicationprotein required for origin-specific DNA binding and interaction with the E2transactivator. Virology 211:385–396.

46. Sedman, J., and A. Stenlund. 1995. Co-operative interaction between theinitiator E1 and the transcriptional activator E2 is required for replicatorspecific DNA replication of bovine papillomavirus in vivo and in vitro.EMBO J. 14:6218–6228.

47. Sedman, J., and A. Stenlund. 1998. The papillomavirus E1 protein forms aDNA-dependent hexameric complex with ATPase and DNA helicase activ-ities. J. Virol. 72:6893–6897.

48. Sedman, T., J. Sedman, and A. Stenlund. 1997. Binding of the E1 and E2proteins to the origin of replication of bovine papillomavirus. J. Virol. 71:2887–2896.

49. Seo, Y.-S., F. Muller, M. Lusky, and J. Hurwitz. 1993. Bovine papillomavirus(BPV)-encoded E1 protein contains multiple activities required for BPVDNA replication. Proc. Natl. Acad. Sci. USA 90:702–706.

50. Seo, Y.-S., F. Muller, M. Lusky, E. Gibbs, H.-Y. Kim, B. Phillips, and J.Hurwitz. 1993. Bovine papilloma virus (BPV)-encoded E2 protein enhancesbinding of E1 to the BPV replication origin. Proc. Natl. Acad. Sci. USA90:2865–2869.

51. Shah, K. V., and P. M. Howley. 1996. Papillomaviruses, p. 2077–2109. InB. N. Fields and D. M. Knipe (ed.), Fields Virology, 3rd ed. Raven Press,New York, N.Y.

52. Stubenrauch, F., and L. A. Laimins. 1999. Human papillomavirus life cycle:active and latent phases. Semin. Cancer Biol. 9:379–386.

53. Sun, Y.-N., J. Z.-J. Lu, and D. J. McCance. 1996. Mapping of HPV-11 E1binding site and determination of other important cis elements for replica-tion of the origin. Virology 216:219–222.

54. Sun, Y., H. Han, and D. J. McCance. 1998. Active domains of humanpapillomavirus type 11 E1 protein for origin replication. J. Gen. Virol.79:1651–1658.

55. Sverdrup, F., and S. A. Khan. 1995. Two E2 binding sites alone are sufficientto function as the minimal origin of replication of human papillomavirus type18 DNA. J. Virol. 69:1319–1323.

56. Sverdrup., F., and G. Myers. 1997. The E1 proteins, p. 37–53. In G. Myers,C. Baker, K. Munger, F. Sverdrup, A. McBride, and H.-U. Bernard (ed.),Human papillomaviruses 1997. Los Alamos National Laboratory, LosAlamos, N.Mex.

57. Thorner, L. K., D. A. Lim, and M. R. Botchan. 1993. DNA-binding domainof bovine papillomavirus type 1 E1 helicase: structural and functional as-pects. J. Virol. 67:6000–6014.

58. Titolo, S., A. Pelletier, A. M. Pulichino, K. Brault, E. Wardrop, P. W. White,M. G. Cordingley, and J. Archambault. 2000. Identification of domains of



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

the human papillomavirus type 11 E1 helicase involved in oligomerizationand binding to the viral origin. J. Virol. 74:7349–73461.

59. Titolo, S., A. Pelletier, F. Sauve, K. Brault, E. Wardrop, P. W. White, A.Amin, M. G. Cordingley, and J. Archambault. 1999. Role of the ATP-binding domain of the human papilloma virus type 11 helicase in E2-depen-dent binding to the origin. J. Virol. 73:5282–5293.

60. Weber, K., and M. Osborn. 1969. The reliability of molecular weight deter-minations by dodecyl sulfate-polyacrylamide gel electrophoresis. J. Biol.Chem. 244:4406–4412.

61. Weisshart, K., P. Taneja, A. Jenne, U. Herbig, D. T. Simmons, and E.Fanning. 1999. Two regions of simian virus 40 T antigen determine cooper-ativity of double-hexamer assembly on the viral origin of DNA replicationand promote hexamer interactions during bidirectional origin DNA unwind-ing. J. Virol. 73:2201–2211.

62. West, M., D. Flanery, K. Woytek, D. Rangasamy, and V. G. Wilson. 2001.Functional mapping of the DNA binding domain of bovine papillomavirusE1 protein. J. Virol. 75:11948–11960.

63. White, P. W., A. Pelletier, K. Brault, S. Titolo, E. Welchner, L. Thauvette, M.Fazekas, M. G. Cordingley, and J. Archambault. 2001. Characterization of

recombinant HPV6 and 11 E1 helicases: effect of ATP on the interaction ofE1 with E2 and mapping of a minimal helicase domain. J. Biol. Chem.276:22426–22438.

64. Woytek, K. J., D. Rangasamy, C. Bazaldua-Hernandez, M. West, and V. G.Wilson. 2001. Effects of mutations within two hydrophilic regions of thebovine papillomavirus type 1 E1 DNA-binding domain on E1-E2 interaction.J. Gen. Virol. 82:2341–2351.

65. Yang, L., R. Li, I. Mohr, R. Clark, and M. R. Botchan. 1991. Activation ofBPV-1 replication in vitro by the transcription factor E2. Nature 353:628–633.

66. Yasugi, T., J. D. Benson, H. Sakai, M. Vidal, and P. M. Howley. 1997.Mapping and characterization of the interaction domains of human papillo-mavirus type 16 E1 and E2 proteins. J. Virol. 71:891–899.

67. Zou, N., J.-S. Liu, S.-R. Kuo, T. R. Broker, and L. T. Chow. 1998. Thecarboxyl-terminal region of the human papillomavirus type 16 E1 proteindetermines E2 protein specificity during DNA replication. J. Virol. 72:3436–3441.

68. Zur Hausen, H., and E. M. de Villiers. 1994. Human papillomaviruses.Annu. Rev. Microbiol. 48:427–447.



http://jvi.asm.org/

Dow

nloaded from

http://jvi.asm.org/

Characterization of the Minimal DNA Binding Domain of the Human Papillomavirus E1 Helicase: Fluorescence Anisotropy Studies and Characterization of a Dimerization-Defective Mutant

Documents