Role of the Hydrophilic Channels of Simian Virus 40 T-Antigen Helicase in DNA Replication

Abstract

The simian virus 40 (SV40) hexameric helicase consists of a central channel and six hydrophilic channels located between adjacent large tier domains within each hexamer. To study the function of the hydrophilic channels in SV40 DNA replication, a series of single-point substitutions were introduced at sites not directly involved in protein-protein contacts. The mutants were characterized biochemically in various ways. All mutants oligomerized normally in the absence of DNA. Interestingly, 8 of the 10 mutants failed to unwind an origin-containing DNA fragment and nine of them were totally unable to support SV40 DNA replication in vitro. The mutants fell into four classes based on their biochemical properties. Class A mutants bound DNA normally and had normal ATPase and helicase activities but failed to unwind origin DNA and support SV40 DNA replication. Class B mutants were compromised in single-stranded DNA and origin DNA binding at low protein concentrations. They were defective in helicase activity and unwinding of the origin and in supporting DNA replication. Class C and D mutants possessed higher-than-normal single-stranded DNA binding activity at low protein concentrations. The class C mutants failed to separate origin DNA and support DNA replication. The class D mutants unwound origin DNA normally but were compromised in their ability to support DNA replication. Taken together, these results suggest that the hydrophilic channels have an active role in the unwinding of SV40 DNA from the origin and the placement of the resulting single strands within the helicase.

Simian virus 40 (SV40) has long been used as the model system to elucidate the mechanism of eukaryotic DNA replication initiation (7, 20, 58). Compared to the multiple origins of replication in eukaryotic cells, a single well-defined origin is utilized in SV40 DNA replication. The 64-bp core origin DNA is composed of three functional regions: a 23-bp perfect palindrome (site II) consisting of four GAGGC pentanucleotides, an imperfect palindrome (EP region), and a 17-bp AT-rich domain (15).

The large tumor antigen (LTag) is a multifunctional protein possessing vital roles in coordinating viral DNA replication and transformation (7, 21, 58, 67). It contains 708 amino acids that fold into several domains (3, 21, 58), including an N-terminal J domain, origin DNA binding domain (OBD), a helicase domain, and a C-terminal domain where host range functions are located (58). So far, the structures of the first three domains have been solved separately using X-ray crystallography and nuclear magnetic resonance techniques. The J domain, representing the first 82 amino acids, shows sequence similarities with DnaJ from Escherichia coli. It interacts with chaperone proteins and is involved in virus replication and transformation (8, 33). The OBD spans residues 131 to 260 and plays an important role in sequence-specific initiation of DNA replication. The recently solved crystal structure of T antigen's OBD reveals that the OBD monomers can form a left-handed spiral hexamer surrounding a positively charged central channel (41). A gap was observed in the spiral giving rise to a “lock wash” structure. The helicase domain of T antigen has been mapped to residues 251 to 627 (22, 24, 34). X-ray crystallography studies of this domain reveal that it contains three smaller structural domains: a zinc domain (D1), a conserved AAA+ helicase motif that binds and hydrolyzes ATP (D2), and a globular domain (D3). Six LTag monomers fold into a hexameric ring structure with a long positively charged central channel, suggesting that the negatively charged DNA interacts directly with the residues in the channel. Six β-hairpins protruding into the central channel are believed to participate in binding and translocating DNA towards the N-terminal OBD of LTag for origin melting and unwinding (24, 57). The hexamer is organized into two tiers with different diameters. Two interfaces are involved in monomer-monomer interactions within a hexamer: the D1-D1 hydrophobic interface in the small tier and the D2-D2 hydrophilic interface in the large tier (34).

In the presence of ATP, small oligomeric forms of LTag assemble into a stable hexamer structure (13, 22, 74). The presence of origin DNA can greatly stimulate the formation of double hexamers (58). Studies of mutant T124A and those in class 4 (40, 72, 78) have suggested that cooperative interactions between two hexamers are needed for origin unwinding. The helicase activity of double hexamers is about 15 times higher than that of single hexamers (54, 64), suggesting, but not proving, that the double hexamer is the active form of the helicase at the origin and at replication forks (2, 74). The double hexamer binds over pairs of pentanucleotides in site II in a head-to-head orientation via their N-terminal OBDs (30, 37, 48, 68). In addition to site-specific binding of the LTag OBD to site II, LTag assembly requires non-sequence-specific contacts between residues in the helicase domain and the flanking sequences (51). Structural distortion of the flanking sequences includes melting of the EP region and untwisting of the AT-rich element (5, 6). It is an early event that does not require ATP hydrolysis but needs ATP binding (6). It may not require double-hexamer formation either, since single LTag hexamers formed on individual pentanucleotides can distort the proximal flanking sequence just like double hexamers (51).

The conformational changes in the flanking sequences are followed by the unwinding of the entire origin, which depends on T antigen's helicase activity. As a processive 3′-to-5′ DNA helicase, LTag is capable of converting the chemical energy of ATP hydrolysis into mechanical work to move the DNA unidirectionally and to separate the duplex strands (25, 75). Extensive mutagenesis studies of LTag showed that efficient origin DNA unwinding relies on a number of subactivities including structural distortion activity (10, 14), ATPase activity (75), and nonspecific single-stranded and double-stranded DNA binding (29, 44, 77).

Single-stranded DNA (ssDNA) loops which serve as templates for further leading- and lagging-strand synthesis are spooled out of the double hexamer by the helicase activity (2, 12, 58, 74). “Rabbit-ear structures” containing T antigen and ssDNA bubbles were observed in several electron microscopy studies (55, 73) supporting this model. Recent nuclear magnetic resonance studies of LTag OBD binding to ssDNA (50), together with previous mutagenesis studies (62), suggested that ssDNA passes over one face of the hexamerized OBD and then threads through the gap present in the open ring structure. Additionally, a cryoelectron microscopy study of the full-length LTag double hexamer assembled at the origin showed low-density masses, putatively assigned to ssDNA, protruding from the junction between the two OBDs of the double hexamer (26). However, how the ssDNAs pass across the helicase domain has not been firmly established (50).

In one DNA pumping model (43), the duplex DNA is pulled into the central channel through the β-hairpin finger structure of the helicase domain, and one strand is threaded to the helicase surface via side channels (24, 26, 34, 50, 57, 70). Based on the crystal structures of the LTag helicase domain obtained under different crystallization conditions, an iris-like motion was proposed to explain how LTag helicase can untwist/distort and melt the origin. In the iris-like motion, the twisting/untwisting motion between the two linked tiers of the hexameric helicase might expand or reduce the dimensions of the central channel in response to the concerted binding of ATP, hydrolysis of ATP, and release of ADP (24, 34, 54).

An alternative helicase model is supported by the recently solved crystal structure of the bovine papillomavirus E1 hexameric helicase encircling ssDNA (19). In this model, the hexamer engages ssDNA through its β-hairpins in the central channel, forming a spiral “staircase” that sequentially tracks the oligonucleotide backbone. The hairpin at the bottom of the “staircase” releases its associated ssDNA phosphate and moves to the top to pick up the next available ssDNA phosphate in response to a sequential (rotary) binding and hydrolysis of ATP. Therefore, only one of the DNA strands is translocated in the hexameric molecule central channel, while the second one is totally displaced to the surface of the helicase domain. The movements within the helicase automatically destabilize the double helix (43, 49). A similar mechanism has also been proposed for prokaryotic helicases, such as DnaB (31, 32) and bacteriophage T7 helicase (1, 18, 27).

At least three other cellular protein factors are loaded on the helicase machine to form a functional replication initiation complex over SV40 origin DNA, including replication protein A (11, 17, 42, 76), topoisomerase I (topoI) (23, 60, 61, 69), and DNA polymerase α/primase (16, 39, 45, 46). The order in which these proteins bind has been studied and determined to be DNA polymerase α/primase and topoI followed by replication protein A (59).

By examining the crystal structure of the helicase domain (34), it is apparent that there are six highly hydrophilic channels that separate neighboring monomers in the hexamer. Many of the residues that line this channel are positioned too far away from the other subunit to participate in monomer-to-monomer contacts (see Fig. 8). Since many of these residues are polar or charged and are conserved among the polyomavirus LTags, we hypothesized that they may interact with SV40 DNA during unwinding from the origin. To investigate this, we performed a mutagenesis study of key residues that line the channel. Mutant proteins were characterized biochemically to determine their ability to oligomerize into hexamers, support SV40 DNA replication, bind and separate origin DNA, hydrolyze ATP, and bind ssDNA. Our data suggest that the hydrophilic channels have a direct role in unwinding the DNA at the origin.

FIG. 8.

Positions of the altered residues in the hydrophilic channel. Ribbon representations of the LTag trimer structure (1SVL) (24) viewed perpendicularly (A and B, left) and parallel (A and B, right) to the hydrophilic channel between two monomers (shown in red and blue). The β-hairpin structure in the central channel is also depicted here. (A) The side chains of residues changed in class A and B mutants are colored in green and yellow, respectively. (B) The side chains of residues changed in class C/C′ and D are colored in silver and cyan, respectively. Images were prepared with Visual Molecular Dynamics, version 1.8.3 (28).

MATERIALS AND METHODS

Oligonucleotide-directed mutagenesis.

Single-point substitutions in SV40 T antigen were performed as described previously (36) by annealing a synthetic oligonucleotide containing a desired point mutation to a uridine-containing single-stranded pSK(−)SVT cDNA (35). The annealed oligonucleotide served as a primer for complementary-strand synthesis by T4 DNA polymerase (New England Biolabs). The newly synthesized strand was then sealed with T4 ligase. The DNA was used to transform Escherichia coli DH5α cells in which the uridine-containing DNA strand was destroyed and replaced with thymidine-containing DNA. Plasmid DNA was prepared from individual colonies and screened for the correct mutation by standard dideoxy DNA sequencing.

Construction of recombinant baculoviruses.

pSK(−)SVT cDNA bearing a point mutation was digested with BamHI, and the smaller DNA fragment was purified and ligated to a BamHI-linearized pVL1393 baculovirus transfer vector (BD Pharmingen). Recombinant baculoviruses were generated by cotransfection of the baculovirus transfer vector and BaculoSapphire (Orbigen) DNA in Spodoptera frugiperda (Sf9) cells as described by the manufacturer. T-antigen-expressing recombinant viruses were detected by immunofluorescence with anti-T-antigen monoclonal antibody PAb101.

Protein purification.

Wild-type (WT) and mutant T antigens were purified from recombinant baculovirus-infected High 5 insect cells by immunoaffinity chromatography with monoclonal antibody PAb101 as previously described (38). The antibody was covalently coupled to CNBr-activated Sepharose 4B beads (Pharmacia) as described in the manufacturer's protocol. The cell lysates were incubated with antibody-containing beads and washed with wash buffer 1 (50 mM Tris, 150 mM NaCl, 1 mM EDTA, 10% glycerol, 1% NP-40 [pH 8.0]) and wash buffer 2 (same as wash buffer 1 but without NP-40). T antigen was then eluted with elution buffer (50% ethylene glycol, 20 mM Tris, 500 mM NaCl, 1 mM EDTA, 10% glycerol [pH 8.5]), dialyzed overnight against dialysis storage buffer (10 mM Tris, 100 mM NaCl, 1 mM EDTA, 50% glycerol, 1 mM dithiothreitol [pH 8.0]), and stored at −20°C. The concentration of the purified protein was determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Coomassie blue staining of the gel by comparison with bovine serum albumin as a standard.

Preparation of DNA substrates.

The 64-bp origin DNA was generated from two complementary oligonucleotides (30). The sequences of the top and bottom strands were as follows: top, 5′-CACTACTTCTGGAATAGCTCAGAGGCCGAGGCGGCCTCGGCCTCTGCATAAATAAAAAAAATTA-3′, and bottom, 5′-TAATTTTTTTTATTTATGCAGAGGCCGAGGCCGCCTCGGCCTCTGAGCTATTCCAGAAGTAGTG-3′.

To prepare the double-stranded DNA substrate, 400 ng of the top-strand oligonucleotide was ³²P labeled at the 5′ end with T4 polynucleotide kinase (New England Biolabs) followed by passage through a Centri-Spin 20 column (Princeton Separations Inc.) to remove unincorporated [γ-³²P]ATP. The radiolabeled top strand was mixed with a 2.5-fold excess of the bottom strand in annealing buffer (50 mM Tris-HCl [pH 8.0] and 10 mM MgCl₂), heated to 95°C, and then allowed to cool down slowly to room temperature. The annealed origin DNA was further purified by applying the reaction mixture to a 7% native acrylamide gel followed by electroelution of the desired DNA fragment from gel slices, extractions with phenol and chloroform, ethanol precipitation, and resuspension in 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA.

Similar procedures were used to label and purify a synthetic replication fork DNA, which was assembled from two partially complementary strands as previously described (56). The sequences of each strand were as follows: top, 5′-TTCTGTGACTACCTGGACGACCGGGTGACTAGCTGCGACGAGATGGGTGCACTGC-3′, and bottom, 5′-GTTCTAGCACTTCGAGTCAACATGGTCGTTCCCGGTGGTCCAGGTAGTCACAGA-3′.

For ssDNA binding, a 55-nucleotide (nt) oligonucleotide corresponding to the bottom strand of the synthetic replication fork (shown above) was end labeled with ³²P and purified.

DNA binding assays.

The ability of T antigen to bind core origin DNA and ssDNA was assessed using gel shift assays. All binding reactions were carried out by incubating 25 ng to 400 ng of T antigen with 2 ng of ³²P-labeled DNA substrate in replication buffer (30 mM HEPES [pH 7.5], 7 mM MgCl₂, 40 mM creatine phosphate, 4 mM ADP, 1 mM dithiothreitol, and 0.1 mg of bovine serum albumin per ml) for 30 min at 37°C. The DNA-protein complexes were then cross-linked with 0.1% glutaraldehyde for an additional 10 min at 37°C and subjected to electrophoresis at 25 mA for 3 h at 3°C on a composite 2.5% acrylamide-0.6% agarose gel in Tris-borate-EDTA buffer (0.089 M Tris, 0.089 M boric acid, 0.02 M Na-EDTA). The gels were dried and exposed to phosphor screens (Molecular Dynamics). Binding activity was quantified by scanning the screens with a PhosphorImager (Molecular Dynamics) and by using ImageQuant 5.0 software.

Origin DNA-unwinding assays.

The same DNA described above in the origin DNA binding assay was also used as a substrate for origin DNA unwinding. Different amounts of T antigen (25 ng or 400 ng) were incubated with end-labeled 64-bp origin DNA at 37°C for 20 min in replication buffer containing ATP. After addition of 5 μl of stop buffer (2% sodium dodecyl sulfate, 0.1% EDTA, 1 mg proteinase K per ml), the reaction mixtures were loaded onto a 7% nondenaturing polyacrylamide gel and subjected to electrophoresis at 110 V for 3 h in Tris-borate-EDTA buffer at 3°C. The gels were dried and exposed to phosphor screens.

Synthetic replication Y-fork-unwinding assays.

Increasing amounts of T antigen (from 25 ng to 400 ng) were incubated with 4 ng ³²P-labeled replication Y fork at 37°C for 30 min in replication buffer with 4 mM ATP in a volume of 20 μl. The reactions were quenched by adding 5 μl of stop buffer (2% sodium dodecyl sulfate, 0.1 M EDTA, and 1 mg of proteinase K per ml). The helicase substrate and released oligonucleotide were separated on a 7% nondenaturing polyacrylamide gel by electrophoresis for 330 V·h at 3°C. The gels were dried, and the DNA bands were visualized and quantitated by autoradiography on a PhosphorImager.

Oligomerization assays.

Four hundred nanograms of T antigen was incubated in the presence or absence of 4 mM ATPγS (Alexis) for 10 min on ice in 30 μl of replication buffer. Proteins were then cross-linked with 0.1% glutaraldehyde for an additional 15 min at 37°C and diluted twofold in Laemmli loading buffer without β-mercaptoethanol or sodium dodecyl sulfate (47). T-antigen oligomers were resolved on a 4 to 20% native gradient gel for 16 h at 100 V and 4°C in 0.05 M Tris (pH 8.8)-0.05 M glycine. The proteins were then blotted onto nitrocellulose membranes and detected by reaction with PAb101 and ChemiGlow buffers (Alpha Innotech) as described previously (23). Blots were then exposed in a FluorChem 8800 instrument.

ATPase assays.

As previously described (78), 600 ng of T antigen was incubated with [γ-³²P]ATP in ATPase buffer (25 mM 1,4-piperazinediethanesulfonic acid [pH 7.0], 0.1 mM NaCl, 5 mM MgCl₂, and 0.01% NP-40) for 1 h at room temperature. The reaction mixtures were spotted on polyethyleneimine-cellulose thin-layer plates (EMD Chemicals Inc.), air dried, and then subjected to ascending chromatography in 0.75 M KH₂PO₄ (pH 3.5). The plates were air dried and exposed to phosphor screens (Molecular Dynamics). The released inorganic phosphate and remaining ATP were then quantitated by autoradiography on a PhosphorImager.

DNA replication assays.

SV40 DNA replication assays were performed with extracts of Ad293 cells as described previously (65). Four hundred nanograms of purified WT or mutant T antigen was used to program a replication reaction mixture containing 350 ng of pSVO1, which contains the origin, in the presence of [α-³²P]dCTP as previously described (65, 66). Labeled DNA was extracted, subjected to a 1.5% agarose gel, and quantitated with a PhosphorImager.

RESULTS

Generation of mutations in the hydrophilic channels.

With the crystal structures of the LTag helicase domain, we identified the amino acid residues in the hydrophilic channels that were far enough away from others so as not to engage in monomer-to-monomer contacts in the hexamer structure (24, 34). Twelve such residues, all polar or charged, were identified, and we generated eight new single-point substitution mutants because we had previously mutated four of these residues (29, 77). In order to maintain the native structure as much as possible, many of the substitutions that we introduced at evolutionarily conserved sites were conservative changes (Table 1; e.g., D484N and T536S), although some less-conserved residues were changed more dramatically (Table 1; e.g., K535T). In this study, we combined our eight new mutants with two others that we previously analyzed (77). Recombinant baculoviruses carrying the genes for mutant T antigens were constructed using oligonucleotide-directed mutagenesis and homologous recombination with BaculoSapphire (Orbigen) DNA. WT and mutant T antigens were expressed in High 5 insect cells and purified with similar yields by immunoaffinity chromatography as described in Materials and Methods.

TABLE 1.

Summary of biochemical properties of single-point mutants^c

Residue and point mutation	Activity (% of WT level) at protein concn (ng)a								Classe
	DNA replication (400)	Unwinding of origin DNA (400)	Helicase (400)	ATPase (600)	Origin binding		ssDNA binding
					25 ng	400 ng	25 ng	400 ng
429,b Asp→Ala	0	4.5 ± 1.0d	84.1 ± 8.4	114.0 ± 0.8	78.1 ± 19.7	100.8 ± 2.9	108.4 ± 21.5	98.6 ± 0.5	A
449, Asn→Ser	0	14.0 ± 9.0d	96.5 ± 2.6	63.7 ± 5.1d	67.6 ± 8.4d	99.1 ± 4.7	102.8 ± 21.5	100.5 ± 4.1	A
515, Asn→Ser	0	9.9 ± 6.9d	95.5 ± 2.3	99.6 ± 1.8	108.0 ± 8.2	104.2 ± 5.6	67.1 ± 15.1d	92.0 ± 8.7	A
460,b Glu→Ala	0	3.1 ± 0.4d	7.7 ± 1.9d	46.7 ± 2.2d	9.3 ± 3.1d	55.2 ± 6.3d	10.1 ± 3.0d	63.8 ± 6.6	B
476, Lys→Arg	0	16.6 ± 8.9d	35.1 ± 2.9d	93.3 ± 6.5	25.6 ± 7.6d	95.2 ± 8.6	46.3 ± 8.7d	94.6 ± 0.8	B
446, Lys→Thr	0	11.0 ± 8.8d	69.6 ± 3.1d	94.5 ± 6.4	90.4 ± 11.4	97.7 ± 6.5	305.9 ± 88.4d	115.1 ± 3.0	C
536, Thr→Ser	0	18.0 ± 10.9d	93.0 ± 2.5	99.4 ± 0.9	113.2 ± 20.3	110.3 ± 17.7	349.1 ± 56.3d	133.5 ± 42.5	C
540, Arg→Lys	0	10.3 ± 7.2d	0	25.9 ± 4.4d	103.1 ± 1.0	100.0 ± 1.1	309.7 ± 55.4d	193.5 ± 65.2	C′
484, Asp→Asn	0	93.9 ± 6.4	105.7 ± 2.7	102.7 ± 1.8	116.2 ± 36.3	101.7 ± 1.2	233.1 ± 27.3d	109.2 ± 1.4	D
535, Lys→Thr	44.7 ± 10.3d	100.6 ± 12.3	101.3 ± 3.9	98.1 ± 3.1	123.8 ± 19.9	102.3 ± 3.7	177.4 ± 61.5d	112.8 ± 5.5	D

^aPercentage of WT activity was determined by comparing the activities of individual mutant LTags with those of the WT at various protein concentrations. Only data for certain concentrations are shown here. Standard deviations were determined from five to seven trials for each assay.

^bPreviously described mutant (77).

^cOligomerization activity at 200 ng of protein was normal for all mutants.

^dSignificant at P = 0.05.

^eMutants were placed in four categories (A, B, C, and D) depending on their activities. The C′ mutant (R540K) is distinguished from the other class C mutants by having lower-than-normal ATPase activity.

Oligomerization activity of mutant T antigens.

It has long been known that for T antigen to perform its helicase function, it first needs to oligomerize into a ring-shaped hexamer structure. In order to ensure that all the mutants assemble into hexamers normally, we carried out oligomerization assays without DNA in the presence or absence of ATPγS and resolved the various structures on 4 to 20% gradient nondenaturing gels (23, 47). As the representative gel in Fig. 1 shows, all 10 mutants displayed fewer monomers and more hexamers in the presence of ATPγS than in its absence (Fig. 1, lanes 3 to 8; Table 1). The oligomerization patterns were similar to that of WT T antigen (Fig. 1, lanes 1 and 2), indicating that the overall conformation of the mutant proteins was not significantly changed and therefore that the normal residues are not likely to be involved in hexamerization, at least in the absence of DNA.

FIG. 1.

Oligomerization assay of WT and mutant T antigens. Four hundred nanograms of immunoaffinity-purified WT or mutant T antigens (N515S, K535T, and T536S) was incubated in the presence (+) or absence (−) of ATPγS. The positions of monomers relative to hexamers in the gel are indicated by the numbers on the left.

DNA replication activity of mutant T antigens.

If a residue within the hydrophilic channels is important for one or more activities required for DNA replication, the mutant protein should be defective in supporting DNA synthesis in vitro. Figure 2 demonstrated that WT T antigen programs the synthesis of replicative intermediates and relaxed circular DNA when added to a cell extract from 293 cells, as previously demonstrated (66). All mutants except one (K535T) showed essentially no activity in supporting DNA replication (Fig. 2; Table 1). K535T displayed about 50% of WT activity. These data provide support for the hypothesis that the residues in the hydrophilic channels participate in DNA replication from the SV40 origin.

FIG. 2.

DNA replication assay of WT and mutant T antigens. Four hundred nanograms of immunoaffinity-purified WT or mutant T antigens was incubated with plasmid DNA containing the SV40 replication origin, cell extracts from 293 cells, replication buffer, and [α-³²P]dCTP. Purified DNA was applied to an agarose gel, and the labeled DNA was visualized and quantitated with a PhosphorImager. The positions of replication intermediates and relaxed circular DNA products are shown.

Origin DNA-unwinding assays.

To determine if the failure of these mutants to support DNA replication could be due to an inability to unwind DNA from the origin, we carried out unwinding assays with an end-labeled 64-bp origin DNA fragment. The amount of released ssDNA was determined from reaction mixtures containing ATP. As shown in Fig. 3 for N515S and N449S, most mutants were completely defective in separating the origin-containing DNA fragment (Fig. 3, lanes 9 to 12; Table 1). Only two mutants (D484N and K535T) (Fig. 3, lanes 5 and 8; Table 1) could unwind the origin normally, and the two residues in question are located close to the external surface of the large tier of the helicase domain (see Fig. 8B). Our interpretation is that the residues in the hydrophilic channels actively participate in DNA unwinding over the origin.

FIG. 3.

Origin DNA unwinding assay of WT and mutant (D484N, K535T, N515S, and N449S) T antigens. Twenty-five or 400 ng of WT or mutant T antigens was incubated with an end-labeled 64-bp origin DNA in the presence of ATP. The DNA was deproteinized and applied to a 7% acrylamide gel. Labeled DNA was visualized and quantitated with a PhosphorImager. The positions of the double-stranded origin DNA and the displaced labeled ssDNA are shown. “Control” and “Boiled” lanes were from reactions with DNA only. The “Boiled” sample was used to show the position of the denatured labeled single strand.

Helicase activity of mutant T antigens.

The failure to unwind the origin normally could be due to a general defect in helicase activity, to an inability to properly recognize the origin and form a double hexameric structure, or to an activity specifically required for origin unwinding. To investigate the first possibility, we carried out a helicase assay using a labeled synthetic replication Y fork in the presence of ATP (Fig. 4; Table 1). The Y fork DNA consisted of a 25-bp duplex region and two 30-nt ssDNA tails. The helicase activity of each mutant at different protein concentrations was determined by comparing the amounts of unwound ssDNA generated to those of WT T antigen. Importantly, five of the eight mutants that showed low origin-unwinding activity displayed normal or near-normal helicase activity (Table 1). Only one mutant (E460A) was severely impaired in helicase activity, and two others (K476R and R540K) were reduced to about a third of normal activity. These deficiencies were apparent at all protein concentrations (Fig. 4 and data not shown). When the ATPase activity of these mutants was checked (Fig. 5; Table 1), E460A and R540K were the only ones that were moderately defective, most likely explaining their lower helicase activity. Together, these data suggest that a number of residues within the hydrophilic channels participate primarily in origin unwinding but are not directly involved in nonspecific helicase or ATPase activity.

FIG. 4.

Synthetic replication Y-fork-unwinding assay of WT and mutant T antigens (N449S and R540K). Increasing amounts of WT or mutant T antigens (as shown) were incubated with a labeled synthetic replication fork in the presence of ATP. Labeled DNA was applied to a 7% acrylamide gel and detected with a PhosphorImager. The positions of the replication fork and unwound single-stranded products are shown.

FIG. 5.

ATPase assays of WT and mutant (N449S, R540K, and N515S) T antigens. Six hundred nanograms of T antigen was incubated with [γ-³²P]ATP in ATPase buffer. Inorganic phosphate was separated from ATP by ascending chromatography and quantitated on a PhosphorImager. The origin and positions of the unhydrolyzed ATP substrate and released iPO₄ are shown.

Origin DNA binding activities of mutant T antigens.

We then asked if the failure to unwind the 64-bp origin-containing fragment might be due to a defect in the formation of a double-hexamer structure over the origin. Binding assays were performed with the ³²P-labeled 64-bp core origin DNA in the presence of ADP. Under these conditions, WT double hexamers readily form over the origin at high protein concentrations (400 ng/reaction) but double and single hexamers form at low concentrations (25 ng/reaction) (Fig. 6, lanes 2 to 4). At 400 ng, all mutants except one possessed a normal ability to form double hexamers over the origin (Fig. 6; Table 1). The exception was E460A, which had about 50% of WT activity. At low protein concentrations (25 ng), E460A was even more defective than WT and another mutant (K476R) demonstrated only about 25% of WT activity. The others retained WT or near-WT levels of origin binding activity. The reduced origin DNA binding by several mutants at low concentrations was not due to a defect in forming hexamers without DNA at low protein concentrations (data not shown). Our interpretation, therefore, is that the mutants' inability to unwind the origin is, in most cases, due not to a problem in binding and forming the proper double hexameric structure nor to a defect in helicase activity.

FIG. 6.

Origin DNA binding assays of WT and mutant (E460A and D484N) T antigens. Increasing amounts (25 ng, 100 ng, and 400 ng) of WT and mutant T antigens were incubated with an end-labeled 64-bp origin DNA fragment in the presence of ADP. Protein complexes were separated from free DNA by electrophoresis on a composite acrylamide-agarose gel. The positions of bound and free DNAs are shown. Sample in lane marked “Control” contained no T antigen.

ssDNA binding activity of mutant T antigens.

One explanation of the data so far is that a number of residues lying within the hydrophilic channels are required for origin-specific unwinding and placement of the resulting single strands in their appropriate sites in and around the helicase domain. One could envision that one strand could thread into one of the six hydrophilic channels after the origin has been unwound. If so, some of the residues in this channel may be necessary for the ability to bind ssDNA. This was tested in an ssDNA binding assay using a 55-mer labeled oligonucleotide in the presence of ATP. We found that at 400 ng of protein, all the mutants except E460A had normal ssDNA binding activity (Fig. 7; Table 1). Interestingly, at low protein concentrations (25 ng per reaction), several mutants exhibited greater-than-normal activity in binding ssDNA (Table 1) and two mutants (E460A and K476R) displayed severely or moderately reduced activity. These same two mutants were also the ones that had a lower-than-normal ability to bind origin DNA at low protein concentrations (Table 1). Our interpretation is, therefore, that the hydrophilic channel does not directly participate in binding ssDNA or that binding of ssDNA by the hydrophilic channels is insignificant compared to the binding by the central β-hairpin-containing channel. The enhanced activity of some mutants may be due to a structural change that facilitates interactions with ssDNA (see Discussion). All of these data are consistent with the idea that many residues within the hydrophilic channels are directly required for unwinding of the origin.

FIG. 7.

ssDNA binding assays of WT and mutant T antigens (E460A and R540K). Increasing amounts, as shown, of WT and mutant T antigens were incubated with a ³²P-end-labeled 55-nt ssDNA in the presence of ATP. The products were separated on a composite gel. The positions of bound and free DNAs are shown. The lane labeled “Control” contained no T antigen.

DISCUSSION

We report in this study that a number of residues that lie in the hydrophilic channels between subunits of the hexameric helicase of T antigen participate in unwinding origin DNA and are therefore required for SV40 DNA replication (Table 1). The structure of the mutant proteins that we studied appeared not to have been seriously changed in structure as gauged by their ability to oligomerize normally. The three-dimensional structure of the T-antigen helicase (24, 34) provides a structural basis for understanding the defects of mutant T antigens. We grouped the mutants into four main classes according to their biochemical activities (Table 1).

Class A.

Mutants D429A, N449S, and N515S were normal or nearly normal in every activity except in unwinding the SV40 origin and in supporting SV40 DNA replication. Since these mutants have normal helicase activity, the WT residues appear to be involved in transforming the double-stranded origin to a partially melted structure. N515 is located on the β-hairpin close to the junction between the main central channel and the hydrophilic channel (Fig. 8A), and N449 is located near the beginning of the hydrophilic channel, but D429 is positioned about two-thirds of the way out of the channel (Fig. 8A, green). There is now strong structural and biochemical evidence that the β-hairpin structure (residues 511 to 517 in LTag) (labeled in Fig. 8) and that of a neighboring α-helix (residues 456 to 462), which lies in the hydrophilic channel but close to the central channel, are directly involved in contacting and separating duplex DNA during helicase action in SV40 LTag and papillomavirus E1 proteins (9, 19, 57). It is therefore possible that N515 and N449 are part of an “active center” for the melting and unwinding reaction. These observations, coupled with the finding that the β-hairpin seems to be positioned over the EP region of the origin (51) where melting is initiated, support our conclusion that some residues in class A participate directly in DNA strand separation.

The third residue in this class (D429) may participate in linking the energy from ATP hydrolysis to DNA unwinding. That residue is adjacent to I428, which moves in different nucleotide states and could serve to eject ADP after the hydrolysis reaction (24). Three other previously characterized mutants (R456A, E460D, and D499A) (29), also altered in the hydrophilic channel, appeared to have properties similar to those of the class A mutants described here. Based on their properties and the location of the altered residue, R456A and E460D may have some commonality with N449S and N515S while D499A may be similar to mutant D429A examined here.

Class B.

Mutants E460A and K476R were defective in many activities including binding ssDNA and origin DNA at low protein concentrations, helicase activity, unwinding of the origin, and supporting DNA replication. One of them (E460A) also had a partial defect in ATPase activity (Table 1). These deficiencies suggest that the mutants are structurally altered although we could not detect a problem with their ability to oligomerize into hexamers in the absence of DNA. Our interpretation is, therefore, that these mutants, although able to form hexameric structures normally, fail to assemble into hexamers and double hexamers in the presence of DNA. For instance, the smear shown in Fig. 6, lanes 5 to 7, may indicate an inability to cooperatively hexamerize with DNA. The placement of these two residues within the hydrophilic channel is shown in Fig. 8A (yellow). It is noteworthy that the same residue (E460), when changed to the conservative Asp, appears to generate a class A mutant (29), whereas a more dramatic change to Ala here results in more severe structural changes.

Class C.

K446T, T536S, and R540K were similar to the class A mutants except that they demonstrated better ssDNA binding activity than did WT T antigen, especially at low protein concentrations. The positions of these three residues in the hydrophilic channel are shown in Fig. 8B. Residue K446 is located close to the central channel, whereas T536 and R540 are on the adjacent subunit about halfway out of the hydrophilic channel and fairly close to the back surface of the large tier (Fig. 8B). How the mutations increase ssDNA binding is not clear. One possibility is that in WT T antigen, the three residues, K446, R540, and T536, interfere in some way with the binding to ssDNA and that this is relieved by mutation. This is possible for R540 and T536 but not K446, since it is deep in the hydrophilic channel (Fig. 8B). Another possibility is that the relationship of the six subunits relative to one another is different in the mutants, thereby permitting better access to ssDNA. For instance, the central channel could be larger or the hydrophilic channel itself could, in the mutants, be more open and attract ssDNA.

In a previous publication (77), we proposed that, based on results obtained with the bacteriophage T7 gene 4 helicase, LTag undergoes an open-filament-to-closed-ring transition when assembly of the hexamer with DNA is completed. Crystallographic studies have shown that the bacteriophage T7 gene 4 helicase domain and E. coli RecA can actually form both rings and filament structures (52, 63, 79). Electron microscopy of SV40 LTag (71) indicates that different structures are possible for that protein as well. In that study, an LTag hexamer was visualized with a clear chiral or pinwheel appearance in the absence of DNA whereas a DNA-associated hexamer had no chirality and appeared ring-like. This suggests that LTag is highly flexible and undergoes a conformational change after the hexamer is fully formed over DNA. These changes point to a third possible explanation for the requirement of the class C residues in origin unwinding, viz., that they participate in the placement of the melted/unwound DNA as the individual subunits assemble to form a hexamer in the presence of ATP. In this model, the residues would be required for a transient association with ssDNA as individual monomers bind during the filament stage of assembly. It's reasonable to think that the increased ssDNA binding activity of the mutants in class C reflects the role of the normal residues. In order for the monomers to stack and move the melted ssDNA, the association with the denatured DNA must be weak. Mutations at these sites appear to enhance stable binding to ssDNA, and this increased interaction leads to a defect in placing the separated strand(s) in their proper places. These ideas do not conflict with models that have been proposed by others for LTag and the papillomavirus E1 helicase (4, 53, 71).

One of the class C mutants (termed C′), R540K, showed a lower ATPase activity (Table 1). In the crystal structure, the positively charged arginine finger R540 interacts with the γ-P_i of ATP directly and participates in ATP hydrolysis (24). Since we changed the arginine to a lysine, our result suggests the importance of the guanidinium group at the end of the R540 side chain in ATP hydrolysis. Because of this, the mutant was abnormal in helicase activity but it bound DNA at least as well as the WT did.

Class D.

The last two mutants in class D (D484N and K535T) were able to unwind origin DNA normally and bind ssDNA slightly better than the WT at low protein concentrations but were defective in supporting DNA replication in vitro. This was especially true of D484N (K535T was about 50% defective). The two residues in question are located near the back surface of the helicase domain (Fig. 8B). The two mutants could be defective in some step subsequent to localized origin melting and unwinding, such as unwinding of circular SV40 DNA (10, 78), or, perhaps less likely, in the association with a cellular protein needed for DNA replication.

Overall, we identified six mutant proteins (those in classes A and C) as being abnormal in the ability to unwind origin DNA and to support DNA replication. At least three other previously characterized mutants (R456A, E460D, and D499A) (29) may also have similar properties. This finding argues strongly that these hydrophilic channel residues are needed for the DNA unwinding from the origin. The normal residues are likely to have different roles in this reaction. For instance, some in class A may be important for the initial unwinding reaction since two of them are located internally, close to the central channel where the strand separation reaction is supposed to occur. Others in this class appear to be needed for coupling ATP hydrolysis to strand separation. Those in class C may be needed for the movement or placement of one or both of the separated strands to their proper positions in the hexamer during assembly over the origin.