Preformed hexamers of SV40 T antigen are active in RNA and origin-DNA unwinding

Abstract

Preformed hexamers of simian virus 40 (SV40) large tumor antigen (T antigen) constitute the bulk of T antigen in infected cells and are stable under physiological conditions. In spite of this they could not be assigned a function in virus replication or transformation. We report that preformed hexamers represent the active T antigen RNA helicase. Monomers and smaller oligomeric forms of T antigen were inactive due to the lack of hexamer formation under RNA unwinding conditions. In contrast to the immunologically related cellular DEAD-box protein p68, the T antigen RNA helicase is found to act in a much more processive way and it does not catalyze rearrangements of structured RNAs. Thereby, it rather seems to resemble other virus-encoded RNA helicases, like vaccinia virus NPH-II. Surprisingly, in our hands preformed hexamers also strikingly bound to and unwound the SV40 replication origin, pointing to a possible role of preformed hexamers in the initiation step of viral DNA replication. Furthermore, we have detected an extra hexamer-specific, high-affinity T antigen ATP binding site with a very slow exchange rate constant, the function of which is discussed.

INTRODUCTION

Simian virus 40 (SV40) large tumor antigen (T antigen) is a multifunctional regulatory protein with numerous biochemical activities (reviewed in 1). It has been classified as a member of superfamily III helicases (2) and can unwind double-stranded (ds) DNA (3) and RNA (4). The properties of the T antigen DNA helicase and its relevance in viral infection have been analyzed extensively (for a recent review, see 5). The role of the RNA helicase so far remains obscure, although a specific function is suggested by several facts: (i) T antigen shares an epitope with a cellular RNA helicase, the DEAD-box protein p68, which is not part of the conserved RNA helicase motifs (6–8); (ii) the T antigen DNA helicase is (d)ATP specific, whereas its RNA helicase is dependent on non-(d)ATP nucleotides (4); and (iii) recently, several DNA and RNA viruses were found to encode RNA helicases (reviewed in 9), some of which have been proven to be essential for viral replication (10,11). In the presence of physiological concentrations of Mg²⁺-ATP, T antigen oligomerizes into stable hexamers (preformed hexamer; 12), the most common form of T antigen occurring in an infected cell (13). Hexamers were shown to represent the active T antigen DNA helicase (14); the structure of the RNA helicase has not been established.

Although T antigen hexamers undoubtedly are the functional unit playing a dual role in the replication of the small circular SV40 DNA genome (see below), it is thought that viral DNA replication is initiated by T antigen monomers, which specifically bind to the palindrome formed by the four GAGGC pentanucleotide sequences of the viral replication origin (‘ori’; reviewed in 15,16). Then, dependent on the presence of Mg²⁺-ATP, bound monomers are assumed to assemble into a double hexamer complex in a cooperative manner. The dodecameric T antigen is bound in a head-to-head configuration to the two halves of the palindrome. As a consequence, the ds structure of the ori-DNA becomes locally distorted (17,18 and references therein). RPA (the eukaryotic single-strand binding protein; 19 and references therein), DNA polymerase α primase (20 and references therein) and possibly additional cellular factors then join the complex to assist the initiation reaction. Eventually, bidirectional DNA replication is carried out by at least 10 cellular proteins (for a recent review, see 21) in addition to the T antigen double hexamer complex, each (hexameric) subunit of which after initiation becomes active as a processive DNA helicase at one of the two forks (14,22,23).

The assumption that SV40 DNA replication is dependent on T antigen monomers which assemble into a functional oligomeric complex after having bound to the ori is mainly based on the observation that isolated hexamers appeared to be inactive in vitro (24,25). However, it is also conceivable that under specific conditions preformed hexamers become functionally inactivated (26), and from a biochemical point of view, it is hard to discern whether hexamer assembly precedes DNA binding or vice versa. Furthermore, under low stringency conditions (low salt), T antigen can efficiently start bidirectional DNA unwinding from non-ori-DNA as well (27). This demonstrates that the structural elements defined for the SV40 ori (T antigen recognition elements 1–4 of binding site II, the early palindrome and the A/T-rich element) are not absolutely required for hexamer assembly, but rather seem to facilitate the loading of the helicase onto the DNA or to be essential for other steps in the initiation process (17,28 and references therein).

The replication of the SV40 genome resembles that of eukaryotic chromosomes in most biochemical aspects and is thus widely used as a model for the respective cellular process (29). Our understanding of how origins are selected in higher eukaryotic cells is far from complete and until now, cellular initiators and replicative DNA helicases could not be identified unambiguously (30–32). From yeast to man, the initiator of DNA replication is thought to be represented by the heterohexameric origin recognition complex (ORC) and/or associated minichromosome maintenance (MCM) proteins. However, in contrast to T antigen or DnaA (the replication initiator protein of Escherichia coli; for a recent review see 33) no duplex DNA opening activity could be demonstrated for this complex in a biochemical assay (34,35 and references therein). The ORC/MCM complex, therefore, either has no such intrinsic activity or it is inactive in its isolated form, thereby possibly resembling preformed T antigen hexamers.

We report here on the biochemical activities of T antigen hexamers, purified under mild conditions from extracts of overproducing insect cells. These preformed hexamers are active in RNA and ori-DNA unwinding, providing evidence that this form of T antigen is capable of biochemical activities required for virus replication.

MATERIALS AND METHODS

Plasmid DNAs

Plasmid pSV-08 is a pUC8 derivative and contains the SV40 control region from nucleotide 5092 to 160, cloned into the SmaI restriction site (for SV40 numbering system see 36). Plasmids pGEM3 and pGEM-7Zf(–) were from Promega, and pGEM-MO1 and pGEM-C4 are pGEM3 derivatives (4). Briefly, pGEM-MO1 contains the HindIII/EcoRI-cloned SV40 control region from 5171 to 41, and pGEM-C4 contains an SV40 DNA fragment from 4459 to 4571 cloned into the same sites.

Purification of T antigen

For isolation of T antigen, Sf9 cells (grown at 27°C in SF900 medium containing 7.5% FCS) were infected with a T antigen-encoding recombinant baculovirus (37) at a MOI of 10 and harvested 40 h post infection. All subsequent steps were performed at 3°C. T antigen was extracted from the cells as described (38) and immunoaffinity purified using monoclonal antibody PAb101, cross-linked to protein A–Sepharose, as described previously (3), except that chelating agents, like EDTA, were omitted from all buffers used. Aliquots of purified T antigen were stored at –70°C.

To purify in vivo preformed T antigen hexamers, Sf9 cell extracts were diluted with 3 vol of buffer A (30 mM triethanolamine–HCl, pH 7.5, 10 mM NaCl, 2 mM MgCl₂, 1 mM ATP, 1 mM DTT, 1% glycerol, 0.02% NP-40) containing 0.2 mg/ml BSA and applied to a histone–agarose (Sigma) column, equilibrated with buffer A plus BSA. After extensive washes with buffer A plus BSA, the column was eluted stepwise with 100, 200 and 350 mM NaCl in buffer A. The 350 mM NaCl-containing fraction was then passed through a DEAE–Sepharose column (0.5 ml) equilibrated with the same buffer (26), and the flow-through was further purified by sucrose gradient centrifugation, performed exactly as described (14), but in the presence of 350 mM NaCl. The 5–7 S fraction contained mostly T antigen monomers, whereas hexamers banded at ∼16 S. Aliquots of the respective fractions were stored at –70°C.

The oligomeric state of T antigen was further analyzed by native gradient polyacrylamide gel electrophoresis. T antigen samples were fixed with glutaraldehyde (0.1% final concentration) on ice for 15 min. After incubation, 10× loading mix (20% Ficoll 4000, 100 mM EDTA, 0.25% bromophenol blue, 0.25% xylene cyanol) was added. The samples were then loaded onto a 4–15% gradient polyacrylamide gel (80:1 acrylamide:bisacrylamide ratio, 250 mM Tris, 500 mM glycine), electrophoresed in 500 mM Tris, 500 mM glycine at 80 V for ∼2.5 h and analyzed by silver staining (39).

SV40 origin-DNA binding assay

A 118 bp DNA fragment containing the SV40 minimal origin (nucleotide 5211–41 of the SV40 genome) was used as substrate in band-shift assays. The DNA fragment was produced by PCR, run with plasmid pGEM-MO1 (0.75 fmol) as a template and a T7 (5′-TAATACGACTCACTA TAGGGAC-3′) and a SV40 (5′-ACGAATTCACTACTT CTGGAATAGCTC-3′) primer (25 pmol each) in the presence of dNTPs (10 µM each) and [α-³²P]dATP (5 × 10⁴ c.p.m./pmol) for 45 cycles at an annealing temperature of 53°C.

T antigen binding reactions, using T antigen (2.5 pmol) monomers or hexamers and the labeled ori-DNA fragment (10 fmol) were conducted at 30°C for 30 min in a total volume of 20 µl containing 40 mM creatine phosphate, 10 mM NaCl, 20 mM triethanolamine–HCl, pH 7.5, 50 ng of a linearized plasmid DNA (4, pGEM-C4; used as a non-sequence-specific competitor), 5 mM MgCl₂, 4 mM ATP, 1 mM DTT, 1% glycerol, 0.02% NP-40 and 4 µg BSA. Nucleoprotein complexes were subsequently fixed with glutaraldehyde (0.1% final concentration) for 5 min at 37°C. Thereafter, 5 µl of gel loading buffer (30% glycerol, 0.3 M TBE, 0.25% bromophenol blue) was added to the reaction mixture. The samples were then loaded onto a 4% polyacrylamide gel (10:1, acrylamide:bisacrylamide) and electrophoresed in 0.5× Tris–borate–EDTA buffer for 90 min (500 V, 20 mA). Dried gels were subjected to autoradiography.

RNA helicase assay

Partially dsRNA substrates were prepared by hybridization of respective complementary RNAs produced by run-off transcription of the linearized plasmids pGEM-C4, pGEM-M01 or pGEM-7Zf(–) exactly as described (4,37). In each case, one strand of a substrate was uniformly labeled with [α-³²P]CTP (specific activity: 2.2 × 10⁵ c.p.m./pmol). RNA helicase activity was assayed as described previously (4). Briefly, reactions were performed in 20 µl volumes containing 2 fmol RNA, 100 mM NaCl, 5 mM MgCl₂, 4 mM ATP, 1.5 mM DTT, 30 mM Tris–HCl, pH 7.5, 30 µg/ml BSA (RNase-free) and 0.5 U/µl RNAsin. After incubation at 37°C for variable periods of time, reactions were stopped by addition of 0.1 vol of 3% SDS, 150 mM EDTA and analyzed by SDS–PAGE and autoradiography. In strand exchange reactions, a 1:3 molar ratio of dsRNA (2 fmol) to a homologous third strand (6 fmol) was used. Reaction conditions were identical to those of the RNA helicase assays.

Ori-DNA-specific DNA helicase assay

In DNA helicase reactions, a 126 bp DNA fragment, containing the SV40 origin (nucleotide 5171–41 of the SV40 genome) was used as a substrate. The DNA fragment was produced by double digestion of plasmid PGEM-MO1 with EcoRI and HindIII and subsequently purified via agarose gel electrophoresis. The purified DNA fragment was ³²P-end-labeled using standard Klenow polymerase (40). DNA helicase reactions, using T antigen (2.5 pmol) monomers or hexamers and the labeled DNA fragment (10 fmol) were conducted at 37°C for 60 min in a total volume of 20 µl containing 10 mM NaCl, 40 mM creatine phosphate, 20 mM triethanolamine–HCl, pH 7.5, 5 mM MgCl₂, 4 mM ATP, 1 µg creatine kinase, 600 ng E.coli single strand binding protein (SSB; Pharmacia) and 1 mM DTT. Helicase reactions were terminated by adding 15 mM EDTA and 0.3% SDS, and DNA products were analyzed by SDS–PAGE and autoradiography.

For electron microscopy, helicase reactions (30 µl) were performed as above, but contained 12.5 pmol of T antigen and 55 fmol of plasmid DNA PSV-08, linearized by BsaI digestion. Reactions were stopped with 0.1% glutaraldehyde (incubated at 37°C for 20 min), spread by the BAC technique and rotary shadowed with tungsten exactly as described (14). Micrographs were taken with a Zeiss 10 transmission electron microscope. For the purpose of calculation of unwinding rates, field views were analyzed for the percentage of partially plus fully unwound DNA molecules.

ATP binding assay

The standard reaction (3 µl) contained 1 pmol of T antigen, 50 mM Tris–HCl, pH 8.25, 0.5 mM magnesium acetate, 0.3 mM EDTA, 20% (v/v) glycerol, 0.1% Triton X-100, 7 mM DTT and 1.3 µM [α-³²P]ATP (1 × 10⁵ c.p.m./pmol). After incubation on ice for 15 min, the solution was filtered through Sephadex G50 (equilibrated in reaction buffer) at 3°C (final volume 20 µl) and fixed with glutaraldehyde (0.1% final concentration) on ice for 10 min. After incubation, 2 µl 10× loading mix (see above) was added. The samples were separated on a 5–15% gradient polyacrylamide gel as described above. The gel was dried and subjected to autoradiography.

[α-³²P]ATP bound to T antigen was also determined by absorption to nitrocellulose membrane filters (Schleicher & Schüll BA 85, 45 µm, 24 mm diameter). After incubation, reactions were diluted to 100 µl with ice-cold wash buffer (50 mM Tris–HCl, pH 8.25, 0.5 mM magnesium acetate, 0.3 mM EDTA, 10 mM (NH₄)₂SO₄, 10% (v/v) glycerol, 0.1% Triton X-100, 5 mM DTT) and filtered through nitrocellulose membranes pre-soaked in wash buffer (8). Filters were washed with 3 ml of ice-cold wash buffer and then with 5 ml of ice-cold wash buffer without Triton and dried. Radioactivity retained on the filters was measured in a liquid scintillation counter. The reaction without T antigen provided a background value of <1 fmol ATP. The K_d of ATP-T antigen was calculated according to Scatchard (41).

RESULTS

Isolation of in vivo formed T antigen hexamers

We have shown previously that T antigen hexamers, but not monomers, are high-affinity binding partners of histones H1 and H3, most probably in order to destabilize nucleosomes in the course of chromatin unwinding during SV40 DNA replication (42). This interaction was employed to gently isolate T antigen hexamers formed in vivo. In order to not deplete T antigen from metal ions, particularly zinc, which seems to be required for maintaining its active protein conformation (26), chelating agents like EDTA were absent from all buffers used in the purification procedure. When applied to histone-affinity chromatography, >50% of total T antigen from an extract of insect cells, infected with a respective recombinant baculovirus, was retained. After elution, this T antigen fraction was further purified by DEAE–Sepharose column chromatography and sucrose gradient centrifugation. Western blot analysis with the T antigen-specific antibody PAb101 of the post-centrifugation fractions revealed that most of the T antigen sedimented at ∼16 S and that this fraction showed a protein homogeneity of >95% (data not shown; compare with 42). Furthermore, the 16 S fraction analyzed by native polyacrylamide gel electrophoresis mainly consisted of hexamers in addition to small amounts of a larger complex, most probably double hexamers (Fig. 1, lane 3). The isolated T antigen hexamers were stable and did not disassemble into smaller oligomerization states or monomers during incubation with Mg²⁺-ATP at 37°C for 30 min. In contrast, T antigen purified from Sf9 extracts by the conventional immunoaffinity chromatography procedure (immunopurified T antigen) consisted of a mixture mainly containing mono-, di- and tetramers as well as some tetra-, penta- and hexamers (Fig. 1, lane 2). Therefore, the elution step from the immunoaffinity column, carried out at pH 10.5, apparently destroyed most T antigen hexamers into smaller complexes and monomers. When immunopurified T antigen was subjected to sucrose gradient sedimentation, a 4–6 S fraction was pooled, which essentially consisted of T antigen monomers plus very low amounts of oligomeric forms (Fig. 1, lane 7). In the activity assays below, this T antigen fraction (referred to as ‘monomers’) was compared with the preformed T antigen hexamers isolated by the histone affinity method. Monomer preparations could be converted almost completely into hexamers by incubation at 37°C for 30 min using Mg²⁺-ATP, but not in the presence of, for example, Mg2⁺-GTP or UTP (Fig. 1, lane 5).

Figure 1.

Native gel electrophoretic analysis of T antigen complexes pre-treated under different conditions. Immunopurified T antigen (12 pmol), fixed with glutaraldehyde, is shown in lane 2. Purified hexamers and a monomer-enriched preparation (referred to as ‘monomers’; see Materials and Methods) of T antigen (6 pmol each) were either fixed with glutaraldehyde directly before electrophoresis (lanes 3 and 7, respectively) or were pre-incubated in the presence of GTP (hexamers, lane 4; monomers, lane 5) or ATP (monomers, lane 6) under RNA helicase reaction conditions at 37°C for 30 min and fixed thereafter. The gel was silver stained after electrophoresis.

T antigen hexamers are active in RNA unwinding

It has been demonstrated previously that T antigen displays an RNA-dependent UTPase/GTPase and an RNA helicase activity (4), the structural basis of which is unknown as yet. Therefore, the different T antigen preparations described above were subjected to RNA helicase activity analysis monitored by gel-shift electrophoresis (43). As seen in Figure 2A, preformed T antigen hexamers showed an efficient RNA helicase activity with a substrate consisting of a partially dsRNA containing a small base-paired region (17 bp; see 4). In contrast, the specific RNA helicase activity of immunopurified protein, which is a mixture of T antigen mono- and oligomers, appeared ∼10-fold lower, but conformed to values reported previously (4). Notably, T antigen monomers were almost inactive (Fig. 2A, lanes 6–8), suggesting that the unwinding activity corresponds to the hexamer content of the respective preparation. RNA unwinding by T antigen hexamers was dependent on a 3′ single-stranded (ss) region of the substrate and required non-(d)ATP nucleotides (e.g. UTP or GTP) as a cofactor (data not shown; see 4). As nucleotides other than (d)ATP do almost not mediate the formation of T antigen hexamers (Fig. 1, lane 5) (33), the RNA unwinding observed (Fig. 2A) must be due to the input preformed hexamers, which proved to be stable under RNA helicase reaction conditions (tested up to 1 h; compare with Fig. 1, lane 4).

Figure 2.

RNA helicase activity analysis of purified hexamers. Helicase activity analysis was monitored by gel-shift electrophoresis and autoradiography. A diagram of the respective substrate used is given to the right of each part with an asterisk marking the labeled strand. The respective heat-denatured substrates (lanes 1) and control reactions without T antigen (lanes 2) were run in parallel. The positions of ds and ssRNAs are marked on the left of each gel. (A) T antigen hexamers constitute the functionally active RNA helicase. Helicase reactions were run with a ³²P-labeled 17 bp ds substrate RNA at the indicated concentrations of T antigen (hexamers, lanes 3–5; monomers, lanes 6–8; immunopurified T antigen, lanes 8–10) at 37°C for 30 min. (B) RNA unwinding by hexamers is not restricted to small dsRNA regions. Helicase reactions were run at the indicated concentrations of purified hexamers and a 123 bp ds substrate RNA as described in (A) in the absence (lane 3) or in the presence of GTP (lanes 4 and 5).

The optimum salt concentration for RNA unwinding by T antigen hexamers was 100 mM NaCl (data not shown); RNA substrates with a ds region of >100 bp (123 bp; Figs 2B and 4B) were unwound efficiently. The length of dsRNA regions, which can be resolved into single strands, reflects the processivity of a respective helicase (43); therefore, the T antigen RNA helicase is a processive enzyme, comparable with vaccinia virus NPH-II, which unwinds dsRNA of similar length (44). Interestingly, RNA unwinding by immunopurified T antigen was more sensitive to salt conditions (data not shown; compare with 4) and showed a lower processivity. This may point to an inhibition of hexamer activity by monomers and/or smaller oligomers.

Figure 4.

Preformed T antigen hexamers efficiently bind to and unwind the SV40 origin of DNA replication. Analyses were monitored by gel-shift electrophoresis and autoradiography. (A) Ori-DNA binding by preformed T antigen hexamers and purified monomers. Reactions with 2.5 pmol of T antigen (preformed hexamers, lanes 2 and 3, or purified monomers, lanes 4 and 5) were performed in the absence (lanes 3 and 5) or in the presence of ATP (lanes 2 and 4). A sample containing ATP, but no T antigen was used as a control (lane 1). The positions of free DNA (ori-DNA) as well as of hexamer (H) and double hexamer (DH) DNA complexes (DNA compl.) are indicated. (B) SV40 ori-DNA unwinding by purified monomers and preformed hexamers in the presence of different (d)NTPs. Ori-DNA-specific helicase reactions were performed without (control) or with 2.5 pmol of T antigen (purified monomers, upper part, or preformed hexamers, lower part) and in the presence of indicated nucleotides (4 mM) at 37°C for 60 min. One reaction mixture was heat denatured at 95°C prior to gel electrophoresis (denatured, lane 2).

Cellular p68, which is immunologically related to T antigen, and also the closely related p72, are low processive RNA helicases, which can unwind RNA double strands of up to ∼35 bp only. In addition to their helicase activity, however, these proteins are capable of catalyzing rearrangements of RNA secondary structures in vitro that resist their helicase activity (43). Such RNA rearrangements proceed via formation and subsequent resolution of RNA branch migration structures, whereby the latter step is dependent on NTP hydrolysis (RNA strand exchange; 43). Interestingly, T antigen hexamers, when tested for such an activity, were completely inactive and did not, for example, replace an RNA strand from a 76 bp partially dsRNA by another homologous one, resulting in a more stable 123 bp RNA duplex (Fig. 3) (42). p68 (data not shown) and p72 (Fig. 3, lanes 4 and 8), in contrast, formed the respective branch migration structure in the absence of ATP and processed it into the expected 123 bp RNA product under conditions that allowed their helicases to be active (especially in the presence of ATP). Instead, the T antigen RNA helicase efficiently unwound the input 76 bp RNA under the strand exchange conditions (Fig. 3, lanes 6 and 7). This result and its inability to induce the formation of the respective branch migration structure (especially in the absence of GTP) states that the T antigen RNA helicase and the p68/p72 helicases clearly differ in their biochemical properties and therefore most probably also in their in vivo functions.

Figure 3.

T antigen hexamers are unable to catalyze RNA strand exchange reactions. The design of the substrate RNAs (with asterisks marking the labeled strand) and the expected course of the strand exchange reaction are shown on top. Thick lines indicate homologous regions of RNA strands. Note that the branch migration structure is a stable reaction intermediate, which needs to be resolved in an NTP-dependent step. Strand exchange reaction mixtures contained the partially ds 76 bp RNA (with one strand ³²P-labeled) plus the respective homologous ssRNA and were incubated without (control, lane 3) or with T antigen (1 pmol, lane 6; 2 pmol, lanes 5 and 7) in the hexameric state in the absence (lane 5) or presence of GTP (lanes 3, 6 and 7) at 37°C for 30 min. As a positive control, recombinant human DEAD-box protein p72 (0.15 pmol) was used, in the absence (lane 4) or presence (lane 7) of ATP instead of GTP. The formation of the respective branch migration structure and the strand exchange activity was monitored by gel-shift electrophoresis and autoradiography. For comparison, the 76 bp dsRNA (lane 1), the ³²P-labeled RNA strand (lane 2) and the 123 bp dsRNA (lane 9) were run in parallel.

Bidirectional unwinding of the SV40 origin of replication by preformed T antigen hexamers

SV40 DNA replication is started by T antigen binding to the minimal viral replication origin (ori). To test whether preformed T antigen hexamers are capable of SV40 ori-DNA binding, we used band-shift assays for the detection of respective DNA–protein complexes formed in vitro. Under the ori-DNA-specific reaction conditions, monomers and hexamers were equally efficient in binding an SV40 ori-DNA fragment in the presence of Mg²⁺-ATP (data not shown). Furthermore, both T antigen forms constituted equal amounts of single and double hexamer–DNA complexes at the same molar protein-to-DNA ratio (250:1; Fig. 4A) (45). However, in contrast to monomers, the ori binding of which clearly was a function of Mg²⁺-ATP, hexamers efficiently bound the ori-DNA independent from the addition of Mg²⁺-ATP. Moreover, the results also clearly indicated that preformed T antigen hexamers could undergo cooperative interactions, which are essential for double hexamer–DNA complex formation as had been shown by mutant studies (46,47).

After having obtained a complex of SV40 ori-DNA and preformed T antigen hexamers, it was of particular interest to study its DNA unwinding activity. Increasing amounts of T antigen hexamers and monomers were comparatively tested for their capacity to unwind a 126 bp dsDNA fragment containing the SV40 origin sequence under the employed specific reaction conditions. Analysis of the reaction products by gel-shift electrophoresis revealed that monomers and hexamers were equally active in DNA unwinding (Fig. 4B lanes 3 and data not shown; see also below) and with both forms of T antigen, 50% of substrate DNA was unwound after a 60 min incubation period at the same molar protein-to-DNA ratio (250:1) at which most of the ori-DNA was converted into the single and double hexamer-bound form in the ori-DNA binding assay (Fig. 4A). Comparable results were obtained when the unwinding activity of T antigen monomers and hexamers was tested with an ori-containing plasmid (pSV-08) (42) and monitored by electron microscopy (Fig. 5) (14,42). The linearized plasmid DNA contained the ori-DNA sequence opposite of the restriction site used for linearization, so that ori-bound T antigen would appear as a spot in the center of the linear dsDNA molecule in an electron micrograph. When incubated with T antigen monomers (data not shown) or preformed hexamers (Fig. 5) in the presence of ATP, we observed bilobed T antigen complexes bound to the ori of almost every DNA molecule. However, hexamers, but not monomers, were capable of ori binding in the absence of ATP as well (data not shown), which confirms the observations above (compare with Fig. 4A). Efficient ATP-dependent unwinding of the plasmid DNA by T antigen monomers and hexamers could be visualized in the presence of E.coli SSB, which stabilizes ssDNA strands (compare with 14,42). The unwinding reaction started from the ori sequence and continued bidirectionally as demonstrated previously (Fig. 5; compare with 14,42). At identical concentrations (molar ratio of T antigen to DNA = 225), T antigen monomers and hexamers unwound 47 and 40% of the DNA molecules, respectively, either completely (seen as DNA single strands covered with SSB) or partially (observed as unwinding intermediates with an unwinding ‘bubble’ and SSB covering ssDNA parts) in a 60 min incubation period. These experiments confirmed that T antigen hexamers are highly effective in ori-DNA unwinding and the slightly lower specific activity measured in comparison to monomers may be due to manipulation during their purification.

Figure 5.

Analysis of the SV40 ori-DNA binding and unwinding activity of preformed T antigen hexamers by electron microscopy. A characteristic field view of an unwinding experiment performed with preformed T antigen (12.5 pmol) hexamers and a linearized (BsaI-cut) plasmid DNA (pSV-08, containing the SV40 origin) is shown. On the left, two dsDNA molecules with ori-DNA-bound T antigen double hexamers and on the right, a partially unwound molecule (lower right) and two DNA single strands (originating from a completely unwound DNA molecule; upper center and right) are seen. The ssDNA is coated with E.coli SSB and therefore condensed more than twice relative to duplex DNA. Identical reactions (not shown) were also performed with purified monomers and without T antigen. Unwinding rates obtained are shown in the lower part of the figure and are expressed as percent of fully and partially unwound DNA molecules obtained after 1 h of incubation. The bar represents 100 nm.

With immunopurified T antigen, it was reported that the (d)NTPs support ori-DNA unwinding with different efficiencies (48). However, the T antigen used in these experiments mainly consisted of monomers due to the immunoaffinity purification procedures used (see Fig. 1), so that the effect of a respective (d)NTP on unwinding efficiency is overlaid significantly by its ability to induce T antigen hexamer assembly. We therefore compared the unwinding of ori-DNA by hexamers and monomers, respectively, in the presence of different (d)NTPs and found that the two forms did not differ from each other with respect to their nucleotide specificity (Fig. 4B). Therefore, our results confirm that only (d)ATP and TTP are capable of efficiently assisting T antigen hexamer DNA unwinding in contrast to RNA unwinding where only GTP or UTP can serve as an energy source. However, the results of this experiment also impressively demonstrate the equivalence of monomers and hexamers with respect to the ori-specific DNA unwinding activity.

Stable binding of ATP to preformed T antigen hexamers

In order to characterize the ATP binding activity of hexamers in detail, we applied a protocol used to label DnaA, the initiator of DNA replication in E.coli (49). Although, in contrast to T antigen, DnaA is not a helicase, ATP has been shown to bind to it with a high affinity (K_s = 0.03 µM), thereby activating its replication function (49). The use of α- instead of [γ-³²P]ATP as a substrate precludes possible phosphorylation events at amino acid side chains which is observed even with highly purified T antigen preparations (50). We found that T antigen hexamers were specifically labeled under the conditions employed (1.3 µM ATP, 3°C; Fig. 6A), which, on the other hand, did not allow monomers to assemble into hexamers (data not shown; compare with 51). Other oligomeric states of T antigen were not capable of ATP binding under these conditions except for a very weak labeling of trimers (Fig. 6A). Neither was the reaction expedited at higher temperatures nor was ATP covalently bound to T antigen as it could be released by SDS treatment (not shown). About one out of five hexamer complexes was able to bind an ATP molecule under labeling conditions with a fast association rate (t_max < 1 min; Fig. 6B). To remove possible pre-bound nucleotides, immunopurified T antigen was incubated at a very high magnesium acetate concentration (3.5 M) as had been employed for G proteins (52). Under these conditions, most hexamers disassembled into T antigen monomers and, therefore, the protein was subsequently (after removal of the salt by gel filtration) incubated at 2 mM Mg²⁺-ATP for 1 h at 30°C to recover hexamer complexes. After purification by sucrose gradient centrifugation (i.e. to remove ATP), the ATP binding capacity of these salt-treated hexamers was analyzed and found to be increased by >5-fold with roughly one ATP molecule being bound per hexamer with a K_s of 0.35 µM (Fig. 6C).

Figure 6.

High-affinity binding of ATP to T antigen hexamers. (A) Label ing of T antigen hexamers. Hexameric (lanes 2 and 4) or immunopurified (lanes 3 and 5) T antigen (0.85 pmol each) was incubated on ice in 3 µl of labeling buffer (see Materials and Methods) for 15 min. After removal of unbound nucleotides by gel filtration, samples were fixed with 0.01% glutaraldehyde and analyzed by native polyacrylamide gel electrophoresis followed by silver staining (lanes 1–3) or autoradiography (lanes 4 and 5). (B) Time course of ATP binding. Reactions with T antigen hexamers (0.85 pmol) were performed for the indicated time periods exactly as described above. [α-³²P]dATP bound to T antigen was determined by absorption to a nitrocellulose membrane, and the radioactivity retained on the filters was measured by liquid scintillation counting. (C) Binding of ATP to salt-stripped T antigen. Immunopurified T antigen was treated with 3.5 M magnesium acetate at room temperature for 30 min, immediately gel-filtrated on a Sephadex G50 column and then incubated with 2 mM Mg²⁺-ATP at 30°C for 1 h. The resulting hexamers were purified by sucrose gradient centrifugation and an aliquot (1.7 pmol of T antigen) was incubated with the indicated concentrations of [α-³²P]dATP as described in (A). [α-³²P]dATP bound to T antigen was determined by absorption to nitrocellulose membrane filters. (D) Bound ATP is not released from T antigen hexamers under DNA helicase reaction conditions. [α-³²P]dATP-labeled T antigen hexamers (A) were incubated under DNA helicase assay buffer conditions in the presence of the indicated concentrations of non- labeled ATP at 20°C for 10 min and analyzed by native polyacrylamide electrophoresis and autoradiography.

Hexamers did not release the bound [³²P]ATP when incubated under DNA helicase reaction conditions in the presence of 3 or 5 mM Mg²⁺-ATP at 20°C (Fig. 6D), although under the same conditions they were found to hydrolyze about 120 molecules of ATP per minute in the presence of poly(dC) (data not shown). Obviously, the high-affinity binding site of T antigen is functionally different from the active ATPase center, and it seems to be blocked at high ATP concentrations where binding of ATP is neither allowed nor can be competed for under these conditions.

DISCUSSION

We have isolated T antigen from recombinant baculovirus-infected insect cells by histone-affinity chromatography and obtained >50% of the bulk of T antigen as purified hexamers. Notably, the proportion of T antigen existing in the hexameric form after immunopurification from the same source is much lower (<20%; this study, but see also 24) and this clearly demonstrates that most hexamers are disrupted during the immunopurification procedure. Interestingly, the fraction of hexamers in immunopurified T antigen has been reported to increase with time of baculovirus infection from <20% (from cells infected for 2 days, like those used in this study) to up to 70% (from cells infected for 4 days; 24), which may be due to increasing protein phosphorylation resulting in more stable but probably inactive T antigen complexes (53). Isolated hexamers are shown here to constitute the T antigen RNA helicase; monomers and smaller oligomers were inactive in the RNA helicase activity assay most probably because they cannot assemble into hexamers in the presence of GTP. The T antigen DNA helicase (14) and RNA helicase activities are therefore carried out by the same hexameric T antigen structure, and the two enzymatic activities probably also share other biochemical properties although their NTP specificities are different. Indeed, RNA unwinding, like DNA unwinding (54), is performed by T antigen in a processive manner and the T antigen RNA helicase seems to resemble that of the vaccinia virus DExH protein NPH-II, which can unwind >100 bp functioning as a molecular motor on RNA (44). Moreover, NPH-II has recently been demonstrated to efficiently catalyze the displacement of the U1A protein from RNA in the course of unwinding (55) and it can be speculated that the T antigen RNA helicase, like its DNA helicase (42), is not repressed by proteins bound to the respective nucleic acid. Thus, both virus-encoded RNA helicases may participate in the structural reorganization of ribonucleoprotein assemblies despite their unknown in vivo functions. The RNA helicase activity of NPH-II has been demonstrated to be essential for virus replication (11), which suggests that the respective T antigen activity could fulfill a similar physiological function, possibly in transcription of viral genes by altering the secondary structure of nascent RNA. However, mutational analyses on T antigen may fail to separate the RNA unwinding and DNA helicase functions as both appear to be performed by the same catalytic unit of the protein (4).

The cellular DEAD-box RNA helicase p68 has been detected on the basis of its immunological relatedness to T antigen and the common epitope of the two proteins might serve as site of interaction with the same target(s) (6). In this context, it is interesting to note that the properties of their RNA helicase activities clearly differ from each other. p68 is able to rearrange structured RNAs (43) whereas T antigen is completely inactive in this respect. T antigen, on the other hand, is a much more processive RNA helicase than p68 (43). The higher processivity of the T antigen RNA helicase may even destroy RNA secondary structures formed by the p68 rearrangement activity. The p68 RNA helicase has been reported to act as an estrogen receptor α co-activator in cooperation with a co-activator RNA (56,57). T antigen, on the other hand, may have a contrary effect on the transcriptional machinery to which it can bind via interaction with CBP (the CREB binding protein) like p68 (58,59). Moreover, p68 has also been shown to function as a cofactor of the chick embryo 5-methylcytosine–DNA glycosylase complex which appears to be involved in tissue differentiation by decreasing the extent of DNA methylation (60), whereas cell transformation by T antigen seems to be linked to a higher degree of DNA methylation (61,62).

Purified T antigen hexamers were also demonstrated here to bind to the SV40 origin of DNA replication efficiently and cooperatively and to initiate bidirectional DNA unwinding. Specifically, it has been excluded that monomers, originating from hexamer complexes disassembled in the course of the assay, account for this activity. First, ori binding of hexamers and formation of double hexamer complexes were observed to occur efficiently in the absence of ATP in contrast to monomers, which were completely inactive under the same conditions. Secondly, T antigen hexamers are stable complexes, as the RNA helicase assay conditions do not allow de novo formation of hexamers as checked by the inability of monomers to unwind RNA (compare with Fig. 2). It is noted that hexamers pre-assembled in vitro from immunopurified T antigen monomers were also active in RNA and ori-DNA unwinding (H.Stahl, unpublished results), contradicting the results published by others (24,25). This discrepancy may be explained by conformational changes with the T antigen hexamer complex in response to particular conditions. In this respect, we observed that the ori unwinding activity of pre-assembled hexamers appears to be very labile and was drastically reduced, e.g. by freezing and thawing, although disassembly of the hexameric structure was not evident under the same conditions (H.Stahl, unpublished results). Furthermore, the use of EDTA in the immunopurification procedure employed by many laboratories (and also in the studies cited above) may have removed divalent cations like zinc, which has been shown to alter the conformation of T antigen (26) and hence may destabilize hexamers not bound to DNA. Taken together, our experiments clearly demonstrate that preformed hexamers efficiently bind to (and unwind) ori-DNA in contrast to previous data showing that only monomers have this capacity (24,25). The potency of preformed hexamers to fulfill the same role in viral DNA replication, which still has to be proven, would alter the current concept on the mechanism of initiation of viral DNA replication by T antigen. Our results suggest that for origin unwinding, T antigen hexamers form prior to their loading onto the DNA, similar to the multi-step mechanism that has been suggested for the loading of T7 helicase-primase protein (gp4A‘) onto its substrate (63). This process may include sequence-specific binding of hexamers by T antigen binding sites in a first step (17), followed by functional double hexamer formation and concomitant melting of respective dsDNA regions. As T antigen hexamers, like gp4A‘ and other hexameric helicases, have a ring-shaped structure, a conformational change, i.e. a ring opening step, is proposed for ssDNA binding in the central channel of the protein.

T antigen hexamers were capable of stable ATP binding. dATP can replace ATP in this reaction (H.Stahl, unpublished results). Monomers as well as T antigen oligomers other than hexamers failed to bind significant amounts of ATP under these conditions. The specificity of the hexamer labeling reaction makes an unspecific attachment of ATP to the T antigen polypeptide chain rather unlikely and points to a regulation of hexamer activities by bound ATP. Binding of [α-³²P]ATP was fast (K_s = 0.35 µM) and obviously did not occur at millimolar ATP concentrations which are essential for efficient T antigen hexamer formation and DNA unwinding. In addition, bound ATP was not released from hexamers under conditions of maximum ATPase. Release of bound nucleotides was observed only under extremely high salt conditions (3.5 M magnesium acetate), which simultaneously disrupt hexamers. From the data, two functionally different types of ATP binding sites within T antigen hexamers can be deduced, which agrees with former observations (64). Further studies will elucidate the loci of the two types of ATP binding sites in the T antigen molecule and the possible hydrolysis of (d)ATP bound to the high-affinity site.

The function of (d)ATP specifically and tightly bound to T antigen hexamers remains obscure at present. The initiator activity of T antigen hexamers, like that of DnaA, may be switched on or off in a way similar to the regulation of G proteins (65,66). Although DnaA is not a helicase, it has an intrinsic ATPase activity, which is used to regulate its initiator function: with ATP bound, it is active, and after hydrolysis of the bound nucleotide, it is inactive in the ADP-bound state. Such a mechanism could be a general way to regulate initiators of DNA replication from E.coli to man. In fact, the (eukaryotic) ORC also binds and hydrolyzes ATP in vitro, and ATP binding to the largest subunit of ORC, Orc1p, stimulates specific binding to ori-DNA (67). Although the exact function of ATP binding to ORC is unknown, it seems to be coupled to the ongoing initiation process. Finally, it is noted that each of the six MCM proteins also contains a putative ATP binding motif, which are required for viability in vivo and coordinated ATP hydrolysis in vitro. Interestingly, mutational analysis has discriminated between two functionally distinct MCM protein subgroups of the heterohexameric complex formed by Mcm4p, 6p and 7p and Mcm2p, 3p and 5p, respectively. The ATP binding motifs of the former are essential for catalysis, and the related motifs in Mcm2p, 3p and 5p serve a regulatory function, whereby specific functional interactions between these two subgroups are required for robust ATP hydrolysis (68). Our observations show parallels between the T antigen hexamer and the MCM complex, which is speculated to represent the replicative DNA helicase in eukaryotic cells (32).

In conclusion, it is obvious that preformed hexamers are a subpopulation of T antigen with distinct biochemical activities. They represent the DNA and RNA helicase (this report), efficiently bind to and unwind the viral replication origin (this report), specifically interact with histones H1 and H3 (42) and express a high-affinity ATP binding site (this report). Although other functions of T antigen, like interaction with DNA polymerase α, apparently are not dependent on the hexameric state (25), it remains to be established whether additional activities of T antigen are also hexamer specific.