JOURNAL OF VIROLOGY, June 2008, p. 5922–5932 0022-538X/08/$08.00⫹0 doi:10.1128/JVI.02723-07 Copyright © 2008, American Society for Microbiology. All Rights Reserved.
Vol. 82, No. 12
Vaccinia Virus DNA Ligase Recruits Cellular Topoisomerase II to Sites of Viral Replication and Assembly䌤 Y.-C. James Lin,1 Jianhong Li,1 Chad R. Irwin,1 Heather Jenkins,1 Luke DeLange,3 and David H. Evans1,2* Department of Medical Microbiology and Immunology1 and Alberta Institute for Viral Immunology,2 Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta T6G 2H7, Canada, and Manitoba Institute of Cell Biology, 675 McDermot Ave., Room ON5027, Winnipeg, Manitoba R3E 0V9, Canada3 Received 21 December 2007/Accepted 3 April 2008
Vaccinia virus replication is inhibited by etoposide and mitoxantrone even though poxviruses do not encode the type II topoisomerases that are the specific targets of these drugs. Furthermore, one can isolate drugresistant virus carrying mutations in the viral DNA ligase and yet the ligase is not known to exhibit sensitivity to these drugs. A yeast two-hybrid screen was used to search for proteins binding to vaccinia ligase, and one of the nine proteins identified comprised a portion (residue 901 to end) of human topoisomerase II. One can prevent the interaction by introducing a C11-to-Y substitution mutation into the N terminus of the ligase bait protein, which is one of the mutations conferring etoposide and mitoxantrone resistance. Coimmunoprecipitation methods showed that the native ligase and a Flag-tagged recombinant protein form complexes with human topoisomerase II␣/ in infected cells and that this interaction can also be disrupted by mutations in the A50R (ligase) gene. Immunofluorescence microscopy showed that both topoisomerase II␣ and II antigens are recruited to cytoplasmic sites of virus replication and that less topoisomerase was recruited to these sites in cells infected with mutant virus than in cells infected with wild-type virus. Immunoelectron microscopy confirmed the presence of topoisomerases II␣/ in virosomes, but the enzyme could not be detected in mature virus particles. We propose that the genetics of etoposide and mitoxantrone resistance can be explained by vaccinia ligase binding to cellular topoisomerase II and recruiting this nuclear enzyme to sites of virus biogenesis. Although other nuclear DNA binding proteins have been detected in virosomes, this appears to be the first demonstration of an enzyme being selectively recruited to sites of poxvirus DNA synthesis and assembly.
fects (7). This suggests that poxvirus topoisomerases serve a biological function resembling that of cellular Topo I. In contrast, poxviruses do not encode type II topoisomerases. The absence of a virus-encoded type II topoisomerase creates a problem, because the complex structures formed during genome replication (19) presumably have to be resolved into monomers, condensed, and packaged into developing particles. The telomere resolution reaction and cleavage of branched DNA structures could be catalyzed by a viral Holliday junction resolvase (6, 11), but neither this enzyme nor the viral topoisomerase (26, 31) is well suited for unknotting the complex structures that are likely formed during virus replication and recombination. One solution could be that poxviruses, like herpesviruses (1), might use cellular type II topoisomerases to disentangle duplex DNAs. Topoisomerases move from cytoplasm to nucleus over the course of the cell cycle and thus might be accessible to viruses replicating in the cytoplasms of infected cells. This hypothesis is supported by the observation that poxvirus growth is inhibited by the topoisomerase II inhibitor etoposide (9, 20), but it does not explain why one can select for viruses exhibiting resistance to this drug and why these viruses carry point mutations in the DNA ligase gene (A50R) (9). Targeted gene deletions create similar, although not identical, phenotypes (9, 27). Interestingly Deng et al. have recently shown that mitoxantrone, a structurally unrelated topoisomerase poison, is also an antipoxviral agent and independently isolated mito-
Topoisomerases play a critical role in modulating DNA superhelical density and in unknotting and decatenating the structures formed during replication, recombination, and repair (reviewed in references 2 and 5). Mammalian cells carry several different kinds of topoisomerases, classified as being either type I or type II enzymes. Both classes of enzyme catalyze DNA strand breakage and rejoining, but type I enzymes catalyze reactions involving single-strand breaks, while the type II topoisomerases catalyze double-strand cleavage reactions. The mammalian type IB enzyme (Topo I) is well adapted for removing the superhelical stress created by the replication and transcription machinery. The two closely related type IIA enzymes (Topo II␣ and Topo II) may also catalyze these same reactions but probably play a more critical role in processes like chromatin reorganization and chromosome segregation. A third class of type IA enzymes (Topo III␣ and Topo III) seems to serve a specialized function in telomere repair (36). Some viruses also encode topoisomerases of which the type IB enzymes encoded by poxviruses are the best characterized (28, 34). Vaccinia viruses lacking the gene encoding this enzyme are still viable but exhibit early-gene-transcription de-
* Corresponding author. Mailing address: Department of Medical Microbiology and Immunology, 141 Medical Sciences Building, University of Alberta, Edmonton, Alberta T6G 2H7, Canada. Phone: (780) 492-2308. Fax: (780) 492-7521. E-mail:
[email protected]. 䌤 Published ahead of print on 16 April 2008. 5922
VOL. 82, 2008
VACCINIA VIRUS DNA LIGASE
xantrone-resistant virus encoding some of the same A50R mutations conferring etoposide resistance (10). The genetics is not explained by targeting of these drugs to the vaccinia ligase, because its activity is not inhibited by etoposide in vitro (33). The A50R gene product, like many other cellular DNA ligases, incorporates additional protein domains that are not essential for catalysis (35). In particular, it carries an ⬃200amino-acid N-terminal domain that is thought to stabilize DNA-protein interactions (32). The N terminus of vaccinia ligase also plays some role in mediating etoposide and mitoxantrone resistance, since drug resistance can be conferred by a single N-terminal C11Y mutation. Mitoxantrone-resistant virus encoding a nearby in-frame indel (adding an isoleucine at position 14) and a mutation in what homologous structures suggest could be the cysteine-11 disulfide partner at cysteine145 have also been isolated (10). In this communication, we show that this complex phenotype is explained by the fact that cellular topoisomerases II␣ and  are recruited to virus “factories” and that the N terminus of vaccinia DNA ligase plays an important role in stabilizing these interactions. Although several nuclear proteins are possibly used by poxviruses to support late-gene transcription (25) and others can interfere with virus replication (12, 23), this appears to be the first demonstration that poxviruses can actively recruit a nuclear replication enzyme to sites of virus biogenesis.
MATERIALS AND METHODS Plasmids, antibodies, and chemicals. The vaccinia virus DNA ligase gene was resynthesized using codons optimized for expression in Escherichia coli (Geneart) and recombined into pDEST17 by using Gateway technology (Invitrogen). The ligase was expressed as an insoluble protein in E. coli, and the purified inclusion bodies were used to produce a polyclonal rabbit antibody (ProSci). Plasmid pSC66-Flag-A50R was constructed by subcloning the modified vaccinia virus ligase gene into vector pSC66 (donated by M. Barry and D. Burshtyn) under the regulation of a poxvirus synthetic early/late promoter. The recombinant protein carries a single Flag epitope (MDYKDDDDK) modifying the native N terminus of vaccinia ligase. A combination of PCR and restriction enzyme digests was used to prepare N- and C-terminal deletions of the Flagtagged gene. Other antibodies were purchased from Stressgen (AK5 anti-topoisomerase II␣/ monoclonal), Sigma (anti-Flag M2), or Li-Cor (IRDye 680 and 800 secondary antibodies). Alexa-488 tagged 2° antibody, rhodamine-phalloidin, and DAPI (4⬘,6⬘-diamidino-2-phenylindole) stains were purchased from Molecular Probes. A monoclonal antibody (MAb) reacting with the vaccinia I3 singlestrand DNA binding protein was prepared commercially (clone 10D11; ProSci), and its properties will be described elsewhere. A 20-mg/ml solution of etoposide was purchased from Novopharm. Infection and transfection. BSC-40 cells were cultured overnight in 100-mm plates and infected with vaccinia at a multiplicity of infection of 5 in phosphatebuffered saline (PBS) at 37°C for 1 h. The inoculums were replaced with fresh optimized minimal essential medium (Gibco) and returned to the incubator for another hour. Lipofectamine reagent was then used as directed by the manufacturer (Invitrogen) to transfect infected cells with the indicated DNAs. BSC-40 cells and vaccinia strain WR were obtained from the American Type Culture Collection. The etoposide-resistant vaccinia strains VP41-8 and ⌬L29 have been described elsewhere (9). The A50R gene in the ⌬L29 virus encodes an in-frame (BamHI-BglII) deletion of amino acid residues 153 to 495, and the VP41-8 virus encodes an A50R C11Y substitution mutation. Where indicated, etoposide was added after the virus inoculums were replaced with fresh medium. Yeast two-hybrid analysis. A yeast two-hybrid assay was used to screen for proteins binding to vaccinia DNA ligase. The wild-type A50R gene was cloned in-frame with the N terminus-encoding region of the GAL4 DNA binding domain in pBG4D2 (TRP1) and the resulting plasmid used to cotransform Saccharomyces cerevisiae strain KGY37 (MAT␣ ade2 trp1⌬-901 leu2⌬-inv::pUC18 his3⌬200 gal4⌬ gal80⌬ ura3⌬inv::GAL1-lacZ lys2⌬-inv::GAL1-HIS3), along with a library of 3.5 ⫻ 106 human brain cDNAs (Clontech) cloned as fusion proteins with the GAL4 activation domain in pACT2 (LEU2). The transformants were
5923
plated on dropout medium and colonies exhibiting His⫹ Leu⫹ Trp⫹ prototrophy and resistance to 1 mM 3-amino-1,2,4-triazole rescreened for -galactosidase activity. The primary screen isolated 28 His⫹ LacZ⫹ clones from 1.2 ⫻ 106 independent transformants. The pACT2-derived plasmids were recovered from His⫹ LacZ⫹ colonies and used to retransform yeast bearing the bait plasmid (pBG4D2) or pBG4D2 encoding the wild-type or mutant (C11Y) ligase::GAL4 fusion protein. Immunoprecipitation assays and Western blotting. BSC-40 cells were cultured in 150-mm dishes and infected with vaccinia strain WR, VP41-8, or ⌬L29 at a multiplicity of infection of 5. At 8 h postinfection, the cells were transferred to 1.5-ml tubes and lysed with 0.5 ml of high-salt buffer (500 mM NaCl, 50 mM Tris-HCl, pH 8.0, 5 mM EDTA, 0.5% NP-40) containing 1⫻-diluted protease inhibitor cocktail (Roche) for 30 min on ice. The cell lysates were centrifuged at 15,000 ⫻ g for 30 min, and the supernatant was recovered and diluted 1:1 with low-salt buffer (50 mM Tris-HCl, pH 8.0, 5 mM EDTA, 0.5% NP-40). The indicated antibodies were added to the extracts to give final concentrations of ⬃10 g/ml and shaken gently overnight at 4°C. The immunocomplexes were recovered using protein G beads (Pharmacia), washed five times with wash buffer (250 mM NaCl, 50 mM Tris · HCl [pH 8], 5 mM EDTA, 0.5% NP-40), and then size fractionated by electrophoresis using an 8% denaturing polyacrylamide gel. Western blots were performed using the indicated primary and secondary antibodies. A Li-Cor Odyssey system was used to image the Western blots. Immunofluorescence microscopy. BSC-40 cells were grown on coverslips in 24-well plates and infected with vaccinia strain WR, VP41-8, or ⌬L29 at a multiplicity of infection of 5. The cells were fixed for 30 min on ice with 4% paraformaldehyde in PBS at the times indicated. The fixed cells were stained with 1:500-diluted anti-topoisomerase II monoclonal (or other) antibodies plus an Alexa 488-conjugated secondary antibody diluted 1:1,000. The cells were counterstained with 10 ng/ml DAPI in PBS and rhodamine-conjugated phalloidin per the manufacturer’s instructions. The specimens were examined using a Zeiss Axioskop fluorescence microscope equipped with DAPI, Alexa 488, and rhodamine filters. Zeiss AxioVision deconvolution software was used to remove out-of-plane images, using the “minimal” setting. Ratiometry measurements were performed as described by Lin et al. (17). Briefly, each of the red-greenblue channels was first converted to a grayscale image. A ratio image was then prepared by dividing each pixel in the image derived from the formerly green channel (topoisomerase II) by the corresponding pixel from the formerly blue channel (DNA). The mean intensity ratio was then calculated for each factory by using the measuring tool provided with ImageJ software (NIH). Immunofluorescence and other plaque assays. BSC-40 cells were seeded into 96-well plates, grown overnight, and then infected for 1 h with ⬃50 PFU per well of each virus. Fresh medium containing 0 to 100 g/ml etoposide was then added, with at least eight wells used for each dose tested. Next day, the cells were fixed for 30 min with ice-cold 4% paraformaldehyde in PBS, washed twice with 0.1 M glycine in PBS with 0.1% Tween 20, blocked with 3% bovine serum albumin in PBS plus Tween for 1 h at room temperature, and then incubated overnight at 4°C with 1/10,000 diluted anti-I3 MAb in PBS plus Tween. The antibody complexes were then stained with an IRDye 800 secondary antibody diluted 1/20,000 and detected using a Li-Cor reader. The plaque sizes and plaque numbers were determined using ImageJ and the 50% effective concentrations (EC50s) calculated using Prism software. The effect of etoposide on virus yield was determined by infecting BSC-40 cells with virus at a multiplicity of infection of 0.1 in six-well plates in the presence of 0 to 100 g/ml etoposide. Next day, the viruses were harvested and their titers determined in triplicate on BSC-40 cells in the absence of drug. Immunoelectron microscopy. BSC-40 cells were infected with vaccinia virus at a multiplicity of infection of 5. At 8 h postinfection, the cells were fixed with 4% paraformaldehyde in PBS for 2 h, dehydrated through an ethanol gradient, and then embedded in LR White resin (EMS). (Only paraformaldehyde was used because topoisomerase antigenicity was destroyed using a glutaraldehyde fixative.) The sections were collected on Formvar-coated grids, stained with MAb directed against topoisomerase ⌱⌱␣/, and counterstained with a secondary antibody conjugated to 10-nm gold particles (EY Laboratories). The specimens were imaged using a Hitachi 7000 transmission electron microscope. Pulsed-field gel electrophoresis and Southern blotting. Pulsed-field gel electrophoresis was performed as previously described (9). Briefly, ⬃106 BSC-40 cells were harvested at different times postinfection and embedded in 0.5% low-melting-point agarose. The agarose plugs were treated with protease (100 g/ml proteinase K, 1% sarkosyl, 0.18 M EDTA, and 10 mM Tris · HCl [pH 7.5]) at 50°C for 24 h, equilibrated in 0.5⫻ Tris-borate-EDTA (TBE) buffer, and then inserted into premolded wells in a 1% agarose gel and sealed in place with 1% low-melting-point agarose. The DNA was separated using a CHEF-DR II apparatus (Bio-Rad) at 6 V/cm and 14°C, using 0.5⫻ TBE buffer with pulse time
5924
LIN ET AL.
J. VIROL.
ramped from 1 to 50 s over 24 h. The separated DNA was depurinated with 0.25 N HCl, denatured and neutralized, transferred to a Zetaprobe membrane (BioRad), and hybridized to BamHI-digested vaccinia genomic DNA labeled with [␣-32P]dCTP, using a random primed labeling kit (Roche). The 32P label was detected using a Typhoon 8600 phosphorimager and processed using ImageQuant. Alternatively, BSC-40 cells were infected with virus at a multiplicity of infection of 10 and treated with or without 80 g/ml etoposide. The DNA was recovered and transferred to a Zetaprobe membrane by using a vacuum manifold, and the virus DNA was then detected by hybridization to a 32P-labeled E9L gene probe.
RESULTS Properties of etoposide-resistant viruses. To confirm that viruses carrying mutations in the vaccinia virus A50R gene exhibit resistance to etoposide, we examined virus growth in the presence of the drug. We used an “in-cell Western” method and a MAb directed against the virus I3 single-strand binding protein to detect the plaques formed after 1 day of growth. (I3 is produced in abundance throughout the virus growth cycle and thus provides a convenient marker of infection.) Increasing amounts of drug caused a reduction in the number of I3-positive plaques without substantially affecting the plaque size (Fig. 1, top). Close inspection of the wells treated with the highest doses of drug showed tiny fluorescent foci, suggesting that vaccinia virus cannot spread beyond a single infected center under such conditions. The plaque counts were used to calculate the EC50, the concentration of drug that reduced the number of plaques to 50% of the control value (Fig. 1, middle). The three different viruses exhibited EC50s (⫾ 95% confidence intervals) of 1.6 ⫾ 0.2, 21 ⫾ 3, and 10 ⫾ 1 g/ml for WR, ⌬L29, and VP41-8 viruses, respectively. Very similar results were obtained by using the amount of I3-dependent fluorescence as a measure of plaque size and by using a “yield reduction assay” where we measured the amount of virus produced after 24 h of growth in the presence of differing concentrations of etoposide (Fig. 1, bottom). We concluded that viruses encoding a 1-kb deletion mutation (⌬L29) were marginally more resistant to etoposide than virus encoding a C11Y substitution mutation (VP41-8), and both were more resistant than virus strain WR. These observations regarding relative resistance of strains WR and VP41-8 are in good agreement with earlier reports, although modern stocks of clinical-grade etoposide may be more active than those historically available (9). The ⌬L29 virus has not been so well characterized, although it has previously been said to exhibit an “intermediate” resistance phenotype compared with other etoposide-resistant strains (9). We also examined the replication properties of these viruses, using orthogonal field gel electrophoresis to measure the production of virus genomes in cells infected with virus at a multiplicity of infection of about 3. These studies showed that the VP41-8 mutant virus replicates as well as does the wild-type virus in the absence of drug (Fig. 2, top). However, the ⌬L29 virus exhibited a replication defect that causes a delay in the appearance of full-length virus genomes and a substantial decrease in the number of monomeric genomes (compared with strain WR) as measured at 24 h postinfection. This replication deficiency is consistent with the small plaque phenotype (Fig. 1, top panel). Treating the infected cells with 80 g/ml (9) etoposide led to the breakage of chromosomal DNA, as judged by the migration of high-molecular-weight DNA into the gel
FIG. 1. Effect of etoposide on vaccinia virus plating efficiency. BSC-40 cells were infected with the indicated vaccinia virus strains in 96-well dishes and then overlaid with fresh medium containing 0 to 100 g/ml etoposide. Next day, the cells were stained with a MAb detecting the I3 single-strand DNA binding protein and visualized with a fluorescent second antibody. The upper panel shows a series of selected fluorescence images, with three wells magnified to show the reductions in WR plating efficiency at higher drug concentrations. Note the cellsized fluorescent spots still visible even at 100 g/ml drug. These were not counted as plaques in subsequent analyses. The middle panel shows a plot of the plating efficiencies for different viruses at different drug concentrations. The lower panel compares the EC50s determined using the plaque counts (plaque reduction; middle panel), the I3 fluorescence (plaque size), or the amounts of virus produced after 24 h of growth in different amounts of drug (yield reduction). Error bars indicate the 95% confidence intervals determined from the curve fits. All three methods show that vaccinia strains carrying mutations in the DNA ligase gene (VP41-8 and ⌬L29) exhibit enhanced resistance to etoposide.
(Fig. 2, upper gel) and inhibited the production of mature virus genomes in all of the infections. As one would predict, the A50R mutations rendered mutant viruses less sensitive to the drug, with more virus DNA produced in mutant infections (compared with the untreated controls) than in cells infected with wild-type virus (Fig. 2, middle panel, lanes 6 and 12). Strain ⌬L29 seemed to exhibit an intermediate level of resis-
VOL. 82, 2008
FIG. 2. Effect of etoposide on vaccinia DNA replication. BSC-40 cells were infected with the indicated strains of vaccinia virus at a multiplicity of infection of 3 and then overlaid with fresh medium containing 0 or 80 g/ml etoposide. The cells were harvested at 5, 10, and 24 h postinfection, embedded in agarose, and digested with protease to mobilize the DNA. The DNAs were fractionated using orthogonal field gel electrophoresis and then visualized by staining with ethidium bromide (upper panel). A Southern blot was used to detect vaccinia virus DNA (middle panel). To determine the relative effects of etoposide on the amount of DNA synthesized 24 h after infection, the amount of radioactivity was determined by phosphorimager analysis and percent DNA synthesis calculated as follows: 100 ⫻ CPM for cells with etoposide at 24 h/CPM for cells without etoposide at 24 h. The ⌬L29 strain exhibits a replication defect not characteristic of VP41-8 (compare lanes 9 and 15 with lane 3), but both mutant strains exhibited greater amounts of DNA synthesis in the presence of etoposide than strain WR (35% and ⬃20% versus 14%). This method might miss small DNA fragments, so we also used a “dot blot” and hybridization procedure to measure the time dependence of virus DNA synthesis in unfractionated cell extracts (bottom panel). The chart shows the percentages of virus DNA formed over 24 h in cells treated with 80 g/ml etoposide, relative to the DNA synthesized (at 24 h) without etoposide (100 ⫻ CPM for cells with etoposide at each time point/CPM for cells without etoposide at 24 h). With DNA synthesis as the endpoint, resistance follows the order VP41-8 ⬎ ⌬L29 ⬎ WR.
tance in this assay, although the concomitant replication defect complicated attempts to accurately measure ⌬L29 viral DNA production in etoposide-treated cells (Fig. 2, lane 18). This issue was further examined using dot blot analysis, which showed that the VP41-8 virus was indeed more resistant to
VACCINIA VIRUS DNA LIGASE
5925
etoposide than strain ⌬L29, as judged by the specific criteria of DNA synthesis. It has been suggested that etoposide treatment also causes the accumulation of higher-molecular-weight forms of virus DNA (9). We noted that etoposide treatment causes the entrapment of high-molecular-weight DNA in the wells of these gels (Fig. 2, upper panel); however, this DNA did not hybridize to our vaccinia DNA probe (despite acid-base treatment prior to transfer) and is presumably cellular in origin. The only abundant virus forms detected in this experiment and at these time points were monomer genomes. Two-hybrid analysis. Because this phenotype is clearly linked to the A50R gene, a yeast two-hybrid genetic screen was performed to look for any human proteins that might interact with vaccinia DNA ligase. A total of nine different cDNA clones that seemed to interact with the wild-type ligase were retrieved from a human brain cDNA library. These clones were characterized by DNA sequencing and BLASTX searches and are listed in Table 1. In order to further limit the number of clones of specific interest, the nine cDNAs were subjected to another two-hybrid screen, using, as “bait,” the mutant A50R gene isolated from strain VP41-8 (9). The C11Y substitution mutation interfered with two-hybrid interactions involving seven of the nine cDNAs (Table 1 and Fig. 3). Interestingly, clone VV46A encoded the C-terminal half, starting at residue 901, of the 1,621-amino-acid  variant of human topoisomerase II (gi 19913408). This region exhibits variable similarity to the related protein topoisomerase II␣ (gi 19913406; 68% amino acid identity from topoisomerase II residues 901 to 1260 and 34% from residue 1261 to end). Because topoisomerase II is the enzyme targeted by etoposide (and, as was later discovered, mitoxantrone), it was decided that this result would be followed up on. A surprising proportion of the other interactions were also affected by the C11Y mutation, suggesting that it might have a substantial effect on the structure of the ligase bait protein. However, the signifi-
TABLE 1. Human cDNAs recovered by yeast two-hybrid screens versus mutant and wild-type vaccinia virus DNA ligase Encoded protein fragment
Topoisomerase II NPC susceptibility protein/ ankyrin repeat protein MIP-T3 (TRAF3 and microtubule binding protein) G-protein signaling modulator 3 R,S-rich mRNA splicing factor SFRS16 RW1 protein/transmembrane protein 131 RING zinc finger protein 692 PHD finger protein 1 (isoform a) Neuronal membraneassociated guanylate kinase
Database identification no.
Two-hybrid interaction Wild type
C11Y
NP_001059 NP_037407
⫹ ⫹
⫺ ⫺
NP_056465
⫹
⫺
NP_071390
⫹
⫺
NM_007056
⫹
⫺
Q92545
⫹
⫺
Q9BU19
⫹
⫺
NP_002627
⫹
⫹
NP_055717
⫹
⫹
5926
LIN ET AL.
J. VIROL.
FIG. 3. Yeast two-hybrid analysis of ligase-topoisomerase interactions. A plasmid encoding a portion of human topoisomerase II was isolated in a yeast two-hybrid screen, using the wild-type vaccinia virus ligase gene (A50R) as “bait.” The LEU2-bearing plasmid encoding the topoisomerase fragment was isolated and used to retransform yeast strain KGY37 along with either the empty TRP1-bearing bait plasmid vector (Vector) or pBG4D2 encoding the mutant (LigC11Y) or wildtype (LigWT) ligase::GAL4 fusion protein. Each transformant was restreaked in duplicate on medium supplemented with (top) or without (bottom) histidine. To maximize the selection for His⫹ prototrophy, 3-amino-1,2,4-triazole (3AT) was also added to the histidine dropout medium.
cance (if any) of these other hits has not been investigated further. Immunoprecipitation analysis. A coimmunoprecipitation strategy was used to confirm the results of the two-hybrid analysis. Recombinant vaccinia virus DNA ligase was prepared in E. coli and the antigen used to raise a polyclonal antibody in rabbits. This antibody (designated B4R1) detected several proteins seen only in virus-infected cells, of which one, exhibiting a mass of ⬃55 kDa, appeared to be vaccinia DNA ligase, as judged by its absence in cells infected with strain ⌬L29 (Fig. 4, top panel) and accumulation late into infection (15; data not shown). We were concerned that the apparent mass falls short of that predicted for vaccinia DNA ligase (63 kDa), but a His10-tagged form of recombinant ligase (32) also exhibits anomalously fast migration on polyacrylamide gels (apparent mass, 57 to 58 kDa), and thus, this appears to be an intrinsic property of the protein. The A50R gene product contains a large proportion of charged residues (14% D plus E and 15% R plus K), which may contribute to the effect. A MAb (AK5) that reacts with both topoisomerase II␣ and II was then used to immunoprecipitate the two proteins from virus-infected BSC-40 cells, along with any bound factors. The precipitated proteins were analyzed by Western blotting using either the AK5 MAb or our polyclonal antibody directed against the A50 protein. The results are shown in Fig. 4. The MAb antibody precipitated and detected two polypeptides from BSC-40 (monkey) cells with masses closely resembling those of human topoisomerases II␣ (174 kDa) and II (183
FIG. 4. Coimmunoprecipitation analysis of ligase-topoisomerase interactions. (Upper panel) BSC-40 cells were cultured in 150-mm dishes and infected (or mock infected) with the indicated virus at a multiplicity of infection of 5. The cells were cultured at 37° for 8 h, lysed with a saltand detergent-containing buffer, and clarified by centrifugation. The topoisomerases were immunoprecipitated with protein G beads and an antibody recognizing both ␣ and  isoforms (MAb AK5) and the immunocomplexes fractionated by denaturing gel electrophoresis. The top and bottom portions of the gel were then Western blotted for either topoisomerase II␣/ (⬃170 kDa) or vaccinia ligase (⬃55 kDa), using MAb AK5 or a polyclonal ligase antibody (B4R1), respectively. The same extracts were also separately blotted for topoisomerase II␣/, vaccinia ligase, and -actin. (Lower panel) BSC-40 cells were infected (or mock infected) with vaccinia virus strain WR and then transfected with plasmid pSC66 encoding just the Flag tag (lane 2) or pSC66 encoding full-length tagged ligase (lane 3), 38- and 19-kDa N-terminal portions of tagged vaccinia ligase (lanes 4 and 5), or a 40-kDa C-terminal portion of the tagged ligase (lane 6). Extracts were prepared at 18 h postinfection, precipitated with MAb to Topo II␣/, fractionated by electrophoresis, and then Western blotted as indicated (HC, heavy chain; LC, light chain). A portion of each extract was also Western blotted to measure ligase expression. The transfection experiment did not resolve the two topoisomerase isoforms, because a shorter electrophoresis time was used to retain the 19-kDa peptide on the gel.
kDa). The antibody directed against vaccinia virus ligase detected an ⬃55-kDa polypeptide that can be precipitated only from cells infected with vaccinia virus strain WR. The mutation in VP41-8 appeared to interfere with the immunoprecipitation
VOL. 82, 2008
of this protein, although the protein is expressed at normal levels in virus-infected cells (Fig. 4, upper panel). None of the 55-kDa polypeptide was detected or precipitated from cells infected with the ⌬L29 virus. As a check on this method, a plasmid encoding a Flagtagged form of the DNA ligase under the control of a synthetic virus early-late promoter was prepared and transfected into vaccinia infected cells and the topoisomerase precipitated with the AK5 MAb. A Western blot was then performed on the precipitated proteins by using a MAb directed against the Flag epitope. The results of this experiment are also shown in Fig. 4 (bottom). The epitope-tagged form of A50 expressed well in virus-infected cells and was also coimmunoprecipitated by the antibody recognizing topoisomerase II␣/. The topoisomerase did not coimmunoprecipitate a Flag-tagged form of ectromelia virus EVM004 protein (data not shown), showing that the Flag epitope was not itself a topoisomerase binding factor. Many different attempts were made to perform reciprocal immunoprecipitation analyses (i.e., immunoprecipitate the DNA ligase and blot for topoisomerase II), using different antibodies and binding conditions, all without success. We also introduced a series of N- and C-terminal deletions into the Flag-tagged ligase and used transient-transfection studies and coimmunoprecipitation analysis to test what portion of the recombinant protein was interacting with topoisomerase II. Our assumption had been that the N terminus of the protein would be critical for this interaction, and indeed, any of the polypeptides encoding the native N terminus of the protein can be coprecipitated with the topoisomerase (Fig. 4, bottom, lanes 3 to 5). The smallest of these peptides still spans both of the cysteines that are mutational targets in drug-resistant viruses (C-11 and C-145), and although fragments this small express rather poorly (which precluded further deletion analysis), this polypeptide was still reproducibly detected in coimmunoprecipitates (Fig. 4, bottom, lane 5). However, a C-terminal protein fragment lacking both C-11 and C-145 (residues 153 to 552) could still also be coimmunoprecipitated with the topoisomerase (Fig. 4, bottom, lane 6). Deletion analysis thus failed to clearly delineate a unique interaction domain. Immunofluorescence microscopy. Immunofluorescence microscopy was used to investigate the behavior of topoisomerase II during the course of virus infection. The results of these experiments are shown in Fig. 5. Topoisomerase II was normally detected only in the nuclei of mock-infected cells (Fig. 5, top row), where it colocalized with DAPI-stained DNA. This association with DNA was most obvious in mitotic cells, such as in the anaphase cell seen in Fig. 5 (top row, second column). The intensities of topoisomerase staining differed greatly from cell to cell, probably due to cells occupying different parts of the growth cycle (16, 30). When the cells were infected with wild-type vaccinia virus, we detected the appearance of DAPIstaining cytoplasmic inclusion bodies, which increased in size and brightness during the first 8 h postinfection and then dissociated to form more-diffuse and less brightly stained structures as virus entered the later stages of infection (the appearance of projectile actin filaments on the cell surface marked late infection). The DAPI-strained structures also contained topoisomerase II antigen. When the cells were infected with strain VP41-8, the virus formed very similar replication “factories,” but the amount of topoisomerase recruited to these
VACCINIA VIRUS DNA LIGASE
5927
sites was notably reduced (Fig. 5; also Fig. 6). The same reduction in topoisomerase staining was also seen in cells infected with the ⌬L29 virus, except that the associated DNA synthesis defect resulted in the production of fewer and smaller viral factories. We also noted that in cells infected with the ⌬L29 virus, there was often some topoisomerase distributed throughout the cytoplasm. We concluded that ligase mutations reduce but never entirely eliminate recruitment of virosome-associated topoisomerase. In addition to these studies, we also infected cells with vaccinia virus strain WR and stained them with the MAb directed against the replicative single-strand DNA binding protein encoded by the vaccinia virus I3L gene. This antibody stained an antigen found only in the DAPI-rich inclusion bodies, confirming that these are sites of virus DNA replication (37). We also examined the distribution of vaccinia virus DNA ligase in cells transfected with Flag-tagged A50R by using a MAb specific for the Flag peptide. Flag-tagged vaccinia virus DNA ligase colocalized within DAPI-strained factories in all of these transfection studies, confirming the work of Kerr et al. with the native protein (13). We concluded that vaccinia virus DNA ligase and cellular topoisomerase II are both recruited to sites of virus replication in the cytoplasms of infected cells. Ratiometric analysis. Subjectively, these studies suggested that less topoisomerase II might be recruited to sites of virus replication in cells infected with virus encoding mutant DNA ligases. To test this hypothesis in a more rigorous manner, we prepared a series of fluorescence micrographs under identical staining and imaging conditions and then used digital imaging tools to measure the ratios of topoisomerase fluorescence to DNA fluorescence within virus factories at 8 h postinfection. Nearly 900 virosomes were analyzed, and the results are shown in Fig. 6. These studies showed that virus encoding the VP41-8 and ⌬L29 mutations recruited significantly less topoisomerase to factories than does wild-type virus (P ⬍ 0.001; NewmanKeuls multiple comparison test). If one assumes that fluorescence is directly proportional to antigen concentration, then about threefold more topoisomerase is recruited to factories in cells infected with a wild-type strain. We also noted that there was a small but still significant difference between the two mutant strains (P ⬍ 0.05), with the VP41-8 virus least able to recruit topoisomerases to virus factories. Thus, a reduction in the amount of topoisomerase recruited to the virosome correlates with the enhanced resistance to etoposide conferred by mutations in vaccinia DNA ligase. Moreover, this effect can be separated from the reduced replicative capacity exhibited by virus carrying deletions within the ligase gene. Transmission electron microscopy. To examine the distribution of topoisomerase antigen in greater detail, we used immunogold labeling and electron microscopy. Cells were infected with vaccinia virus strain WR and then fixed and sectioned at 8 h postinfection. Topoisomerase antigens were detected using the AK5 MAb plus a secondary antibody linked to 10-nm gold particles. We observed that gold particles were distributed throughout the nucleus in mock-infected cells, with little or no antigen detectible in the cytoplasm (data not shown). This changed in infected cells, where the cytoplasm is seen to accumulate the electron-dense fibrous structures characteristic of classical viral factories (8). These structures include the viroplasm and associated crescent structures and
5928
LIN ET AL.
J. VIROL.
FIG. 5. Immunofluorescence microscopy. BSC-40 cells were cultured on glass slides and then infected (or mock infected) with the indicated virus strains at a multiplicity of infection of 5. At times ranging from 0 to 18 h, the cells were fixed and stained with fluorescent probes for DNA (DAPI; blue), topoisomerase (AK5 1° MAb plus Alexa 488 conjugated 2° antibody; green), and actin (rhodamine-phalloidin; red). MAbs directed against the I3 vaccinia single-strand DNA binding protein (␣SSB; clone 10D11) and transfected FLAG-tagged vaccinia ligase (M2 1° MAb) plus an Alexa 488-conjugated 2° antibody (green) were also used to differentiate virosomal from nuclear DNA and confirm the recruitment of vaccinia ligase to these structures (bottom panel). All images were captured at the same exposure settings (20, 200, and 400 ms for blue, green, and red channels, respectively).
VOL. 82, 2008
VACCINIA VIRUS DNA LIGASE
5929
this context, our analysis showed that vaccinia strain WR recruited both Topo II␣ (MAb 4A12) and Topo II (MAb 3B6) to sites of virus replication, whereas little or no topoisomerase antigen was detected in the factories formed by mutant viruses. After examining hundreds of cells stained with antibodies 4A12 and 3B6 at 8 h postinfection, we also noticed some subtle differences in the distributions of topoisomerases II␣ and II before and after the disassembly of cytoplasmic factories. Although it is a difficult observation to quantify, it appeared that Topo II␣ was more strongly associated with DAPI-stained factories before their dispersion, whereas Topo II was more closely associated with viral DNA after the viroplasm had started to dissociate. DISCUSSION
FIG. 6. Ratiometric analysis of immunofluorescence images. BSC-40 cells were cultured on glass slides and then infected with the indicated strains of vaccinia virus at a multiplicity of infection of 5. At 8 h postinfection, the cells were stained for DNA (DAPI) and topoisomerase II␣/ (AK5 1° MAb plus Alexa 488-conjugated 2° antibody) and a series of fluorescence images acquired all at the same excitation and acquisition settings. The blue and green channels were converted to grayscale images, and the ratio of topoisomerase to DNA integrated across the area of each virosome by using ImageJ software. The upper panel illustrates an example of one image, and about 900 factories were sampled to assemble the data shown in the lower panel. The figure shows the results of a New-Keuls multiple comparison test, and a one-way analysis of variance also indicates a significant difference among the three strains (P ⬍ 0.0001).
immature virions as well as newly assembled intracellular mature virus (IMV) (Fig. 7). Interestingly, the immunogold label was mostly localized within the viroplasm, with some labeling also seen near or within immature virus forms and in the nucleus. Little or no topoisomerase antigen was detected in the surrounding cytoplasm or in IMV. We also used a sucrose gradient to further purify the IMV and then Western blots to screen the fractions for both topoisomerase and virus protein p28. No topoisomerase II was detected in any of the fractions containing an abundance of the p28 core antigen (data not shown). Thus, we can show that the topoisomerase II is recruited to sites of virus genome replication and capsid assembly but can find no evidence to support the hypothesis that it is incorporated into mature forms of assembled virus. Topoisomerase specificity. The AK5 antibody used in all of the preceding studies recognizes both topoisomerase II␣ and II isoforms, but the two-hybrid analysis originally detected an interaction between vaccinia ligase and topoisomerase II. Are the interactions we detected specific for either isoform? To address this question, we used immunofluorescence microscopy and MAbs (kindly provided by A. Kikuchi), which can differentiate between the two proteins (30). The results are shown in Fig. 8. These antibodies detected some differences in the behavior of the two proteins, with Topo II␣ exhibiting a stronger and more variable pattern of nuclear staining than Topo II. Virus infection also appeared to decrease the amount of nuclear Topo II, regardless of genotype. Within
Nearly all poxviruses encode DNA ligases. These enzymes are not essential for virus growth in culture (4, 14), but mutant virus lacking a catalytically active enzyme (like the ⌬L29 strain) exhibit defects in DNA synthesis in many cell types (Fig. 2) (27). A peculiar property of these mutant strains is that ligase mutations also create resistance to etoposide and mitoxantrone, even though etoposide has no known effect on the activity of purified poxvirus DNA ligases (9, 10, 33). Both drugs are highly specific topoisomerase poisons, which work by trapping topoisomerase cleavage complexes on DNA and thus blocking processes like DNA replication and repair (3). They are not expected to affect the activity of DNA ligases. Our data suggest that drug-resistant virus acquires that phenotype when the A50R mutations reduce the amount of topoisomerase recruited to the virosomes, and this reduces the number of cleavage complexes that might otherwise be trapped on virus DNA in the presence of these drugs. What effect this has on virus biology depends upon the A50R mutation. If one uses assays that measure many cumulative elements of the infection cycle, including maturation and spread, then the ⌬L29 (null) mutation engenders a marginally greater degree of etoposide resistance (Fig. 1). However, DNA replication is most advantaged by the VP41-8 (point) mutation (Fig. 2). Three different types of experiment provide evidence for an interaction between vaccinia ligase and cellular type II topoisomerases. First, a yeast two-hybrid screen retrieved a portion of topoisomerase II from a human cDNA library, using, as bait, full-length vaccinia DNA ligase (Fig. 3). One can also interfere with this interaction by introducing a C11Y mutation into A50R. This mutation confers resistance to topoisomerase inhibitors, while presumably having little effect on ligase catalytic activity, as judged by the location of the mutation and the relatively normal levels of DNA synthesis seen in cells infected with VP41-8 virus (Fig. 2). These data also suggest that the interaction region involves (in part) the protease-resistant N terminus of vaccinia ligase (32) and the C-terminal portion of human topoisomerase II (residue 901 to end in the Topo II variant). The recovery of a Topo II clone from a fetal brain cDNA library is not unexpected. Topo II plays an important role in neuronal development and is also the predominant isoform expressed in nonproliferating tissues (18). Immunoprecipitation studies provided additional evidence for an interaction between the two proteins. Native vaccinia DNA ligase was coimmunoprecipitated with topoisomerase
5930
LIN ET AL.
J. VIROL.
FIG. 7. Immunoelectron microscopy. BSC-40 cells were infected with vaccinia strain WR at a multiplicity if infection of 5 and cultured for 8 h. The cells were fixed and embedded in resin, sectioned, stained with MAb to topoisomerase II␣/ (〈⌲5) plus a secondary antibody conjugated to 10-nm gold particles, and then imaged by transmission electron microscopy. Note the gold label associated with virosomal elements and immature virions (IV) and the apparent absence of topoisomerase antigen in IMV.
II␣/ by using a MAb recognizing both proteins (Fig. 4). Moreover, this interaction was not detected in cells infected with etoposide-resistant mutant virus. Controls showed that normal levels of ligase polypeptide were still expressed in cells infected
with the VP41-8 virus, suggesting that the phenotype is not caused by the point mutation destabilizing the protein and promoting its degradation. A Flag-tagged version of vaccinia ligase was also coimmunoprecipitated with the MAb directed
FIG. 8. Immunofluorescence analysis of topoisomerase II␣ and II recruitment specificity. BSC-40 cells were cultured on glass slides and then infected (or mock infected) with the indicated virus strains at a multiplicity of infection of 5. Eight hours after infection, the cells were fixed and stained with fluorescent probes for DNA (DAPI; blue), Topo II␣ (4A12 1° MAb) or Topo II (3B6 1° MAb) plus an Alexa 488-conjugated 2° antibody (green), and actin (rhodamine-phalloidin; red). All images were captured at the same exposure settings (20, 200, and 400 ms for blue, green, and red channels, respectively). Both enzyme isoforms are recruited to virosomes in cells infected with wild-type virus. The white arrows help indicate the small virosomes typically seen in cells infected with ⌬L29 virus.
VOL. 82, 2008
against topoisomerase II␣/. Despite many attempts, we were never able to perform the reciprocal coimmunoprecipitation reaction. We have noted that topoisomerase II is a very labile protein in infected-cell extracts and probably much less abundant than the viral DNA ligase. Thus, one possible explanation for this difficulty is that only a small fraction of the total ligase protein forms stable complexes with the topoisomerase. Alternatively, the epitope(s) targeted by two different antibodies is occluded by the topoisomerase interaction. Immunofluorescence microscopy provided a third, and more direct, form of evidence for topoisomerase II accumulation within sites of virus replication. In uninfected BSC-40 cells, the nucleus was the only site where topoisomerase II antigens can be easily detected (Fig. 5). This is consistent with other reports suggesting that, while topoisomerases may shuttle between cytoplasm and nucleus, the nucleus is the primary site of accumulation in nonmitotic cells (21). The distribution of topoisomerase II antigens changes in poxvirus-infected cells. Soon after infection, DAPI-binding structures are seen in the cytoplasm and can be shown to be sites of virus replication because they contain vaccinia single-strand DNA binding protein and DNA ligase (Fig. 5) as well as virus DNA detectible using fluorescence in situ hybridization analysis (J. Li and D. H. Evans, unpublished data). These structures also contained topoisomerase II (Fig. 5), with no obvious preference for topoisomerase II␣ or II (Fig. 7 and 8). Most importantly, less topoisomerase II was recruited to virosomes in cells infected with ⌬L29 and VP41-8 viruses even though normal levels of DNA synthesis were detected in cells infected with strain VP41-8 (Fig. 5 and 7). Ratiometric analysis suggested that the ligase mutations reduced the amount of virosome-associated topoisomerase approximately threefold, with the point mutation in strain VP41-8 having a slightly greater effect than the deletion encoded by strain ⌬L29 (Fig. 6). The greater amounts of DNA synthesis seen in cells infected with strain VP41-8 might explain this effect, since this increase would reduce the topoisomerase-to-DNA fluorescence ratio. It is also consistent with the reduced impact of etoposide on DNA replication in VP41-8 versus ⌬L29-infected cells (Fig. 2). What role might topoisomerase II serve in vaccinia virus replication and/or assembly? Immunofluorescence microscopy shows that topoisomerase antigens can be detected in factories as early as 2 h postinfection (data not shown) and stay associated with virus DNA throughout most of the infectious cycle (Fig. 5). However, the topoisomerase-ligase interaction can be disrupted with a point mutation (Fig. 3 and 4) without greatly affecting the amount of DNA synthesis in untreated cells (Fig. 2, compare lanes 1 to 3 with lanes 7 to 9). This suggests that the recruitment of topoisomerase II to the virosomes is probably not essential for DNA replication, perhaps because its activity can be complemented in part by the viral type I enzyme. The topoisomerase is dispersed along with virus DNA late in infection (Fig. 5). Immunoelectron microscopy also detects an abundance of antigen located within virosomes and within (or near) maturing virus particles (Fig. 7). However, we could not detect the topoisomerase in IMV either by immunoelectron microscopy (Fig. 7) or by Western blotting of purified particles. Nor has its presence ever been detected in virions by using mass spectrometry. Although one cannot exclude the possibility that topoisomerase antigens are simply rare and/or some-
VACCINIA VIRUS DNA LIGASE
5931
how obscured during packaging, the simplest interpretation of these data is that topoisomerase II is not packaged into mature virus. This might be easily explained if the amount of bound ligase regulated the amount of associated topoisomerase. Nicks would provide binding sites for the two enzymes on newly synthesized DNA (where the topoisomerase might serve to unknot and decatenate the genome), and the loss of nicks through their ligation would cause the simultaneous dissociation of both enzymes. Adding etoposide or mitoxantrone to this environment would cause the conversion of a presumably beneficial enzymatic association into a toxic adduct that inhibits virus replication and/or maturation (10). An interesting feature of what poxviruses are doing is that ligase I and topoisomerase II are both also recruited to cellular sites of replication. DNA ligase is recruited through an interaction with PCNA and topoisomerase II either independently or perhaps also through an interaction with PCNA (22, 24). Vaccinia would not require such a cofactor to do this, because the special nick-sensing capabilities of vaccinia ligase can provide an alternate DNA targeting and stabilizing strategy. Cells and virus could thus both be regulating topoisomerase II activity at sites of repair and replication, but by using different methods. Based primarily upon studies using enucleated cells (29), it has long been thought that few, if any, of the replication enzymes sequestered in the nuclei of cells are required for cytoplasmic poxvirus DNA replication. However, the results of our current investigations suggest that a reappraisal of this hypothesis might be required. In recent years, it has become clear that many nuclear DNA binding proteins are attracted to sites of viral replication (YY1 [25], BAF [23], and HMG20A [12]), and now, we have shown that vaccinia also recruits the topoisomerase II␣/ enzymes through a specific interaction with the virus-encoded ligase. This leads one to wonder what other nuclear functions might be required for poxvirus biogenesis and what process leads these proteins to exit the nucleus. ACKNOWLEDGMENTS We thank M. Barry, D. Burshtyn, D. Gietz, D. Hruby, A. Kikuchi, and S. Shuman for the gift of reagents and helpful advice and R. Sherborne for help with electron microscopy. This work was funded by an operating grant from the CIHR (MOP 10923) to D.H.E. REFERENCES 1. Advani, S. J., R. R. Weichselbaum, and B. Roizman. 2003. Herpes simplex virus 1 activates cdc2 to recruit topoisomerase II alpha for post-DNA synthesis expression of late genes. Proc. Natl. Acad. Sci. USA 100:4825–4830. 2. Champoux, J. J. 2001. DNA topoisomerases: structure, function, and mechanism. Annu. Rev. Biochem. 70:369–413. 3. Chen, G. L., L. Yang, T. C. Rowe, B. D. Halligan, K. M. Tewey, and L. F. Liu. 1984. Nonintercalative antitumor drugs interfere with the breakage-reunion reaction of mammalian DNA topoisomerase II. J. Biol. Chem. 259:13560– 13566. 4. Colinas, R. J., S. J. Goebel, S. W. Davis, G. P. Johnson, E. K. Norton, and E. Paoletti. 1990. A DNA ligase gene in the Copenhagen strain of vaccinia virus is nonessential for viral replication and recombination. Virology 179: 267–275. 5. Corbett, K. D., and J. M. Berger. 2004. Structure, molecular mechanisms, and evolutionary relationships in DNA topoisomerases. Annu. Rev. Biophys. Biomol. Struct. 33:95–118. 6. Culyba, M. J., N. Minkah, Y. Hwang, O. M. Benhamou, and F. D. Bushman. 2007. DNA branch nuclease activity of vaccinia A22 resolvase. J. Biol. Chem. 282:34644–34652. 7. Da Fonseca, F., and B. Moss. 2003. Poxvirus DNA topoisomerase knockout mutant exhibits decreased infectivity associated with reduced early transcription. Proc. Natl. Acad. Sci. USA 100:11291–11296. 8. Dales, S., and L. Siminovitch. 1961. The development of vaccinia virus in
5932
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19. 20. 21.
22.
LIN ET AL.
Earle’s L strain cells as examined by electron microscopy. J. Biophys. Biochem. Cytol. 10:475–503. DeLange, A. M., M. S. Carpenter, J. Choy, and V. E. Newsway. 1995. An etoposide-induced block in vaccinia virus telomere resolution is dependent on the virus-encoded DNA ligase. J. Virol. 69:2082–2091. Deng, L., P. Dai, A. Ciro, D. F. Smee, H. Djaballah, and S. Shuman. 2007. Identification of novel antipoxviral agents: mitoxantrone inhibits vaccinia virus replication by blocking virion assembly. J. Virol. 81:13392–13402. Garcia, A. D., and B. Moss. 2001. Repression of vaccinia virus Holliday junction resolvase inhibits processing of viral DNA into unit-length genomes. J. Virol. 75:6460–6471. Hsiao, J. C., C. C. Chao, M. J. Young, Y. T. Chang, E. C. Cho, and W. Chang. 2006. A poxvirus host range protein, CP77, binds to a cellular protein, HMG20A, and regulates its dissociation from the vaccinia virus genome in CHO-K1 cells. J. Virol. 80:7714–7728. Kerr, S. M., L. H. Johnston, M. Odell, S. A. Duncan, K. M. Law, and G. L. Smith. 1991. Vaccinia DNA ligase complements Saccharomyces cerevisiae cdc9, localizes in cytoplasmic factories and affects virulence and virus sensitivity to DNA damaging agents. EMBO J. 10:4343–4350. Kerr, S. M., and G. L. Smith. 1991. Vaccinia virus DNA ligase is nonessential for virus replication: recovery of plasmids from virus-infected cells. Virology 180:625–632. Kerr, S. M., and G. L. Smith. 1989. Vaccinia virus encodes a polypeptide with DNA ligase activity. Nucleic Acids Res. 17:9039–9050. Kimura, K., M. Saijo, M. Ui, and T. Enomoto. 1994. Growth state- and cell cycle-dependent fluctuation in the expression of two forms of DNA topoisomerase II and possible specific modification of the higher molecular weight form in the M phase. J. Biol. Chem. 269:1173–1176. Lin, Y. C., N. G. Grigoriev, and A. N. Spencer. 2000. Wound healing in jellyfish striated muscle involves rapid switching between two modes of cell motility and a change in the source of regulatory calcium. Dev. Biol. 225: 87–100. Lyu, Y. L., and J. C. Wang. 2003. Aberrant lamination in the cerebral cortex of mouse embryos lacking DNA topoisomerase IIbeta. Proc. Natl. Acad. Sci. USA 100:7123–7128. Merchlinsky, M., and B. Moss. 1989. Resolution of vaccinia virus DNA concatemer junctions requires late-gene expression. J. Virol. 63:1595–1603. Meresse, P., E. Dechaux, C. Monneret, and E. Bertounesque. 2004. Etoposide: discovery and medicinal chemistry. Curr. Med. Chem. 11:2443–2466. Mirski, S. E., K. E. Sparks, B. Friedrich, M. Kohler, Y. Y. Mo, W. T. Beck, and S. P. Cole. 2007. Topoisomerase II binds importin alpha isoforms and exportin/CRM1 but does not shuttle between the nucleus and cytoplasm in proliferating cells. Exp. Cell Res. 313:627–637. Montecucco, A., R. Rossi, D. S. Levin, R. Gary, M. S. Park, T. A. Motycka, G. Ciarrocchi, A. Villa, G. Biamonti, and A. E. Tomkinson. 1998. DNA ligase
J. VIROL.
23.
24.
25. 26.
27.
28. 29. 30.
31.
32. 33. 34. 35.
36.
37.
I is recruited to sites of DNA replication by an interaction with proliferating cell nuclear antigen: identification of a common targeting mechanism for the assembly of replication factories. EMBO J. 17:3786–3795. Nichols, R. J., M. S. Wiebe, and P. Traktman. 2006. The vaccinia-related kinases phosphorylate the N⬘ terminus of BAF, regulating its interaction with DNA and its retention in the nucleus. Mol. Biol. Cell 17:2451–2464. Niimi, A., N. Suka, M. Harata, A. Kikuchi, and S. Mizuno. 2001. Colocalization of chicken DNA topoisomerase IIalpha, but not beta, with sites of DNA replication and possible involvement of a C-terminal region of alpha through its binding to PCNA. Chromosoma 110:102–114. Oh, J., and S. S. Broyles. 2005. Host cell nuclear proteins are recruited to cytoplasmic vaccinia virus replication complexes. J. Virol. 79:12852–12860. Palaniyar, N., E. Gerasimopoulos, and D. H. Evans. 1999. Shope fibroma virus DNA topoisomerase catalyses holliday junction resolution and hairpin formation in vitro. J. Mol. Biol. 287:9–20. Parks, R. J., C. Winchcombe-Forhan, A. M. DeLange, X. Xing, and D. H. Evans. 1998. DNA ligase gene disruptions can depress viral growth and replication in poxvirus-infected cells. Virus Res. 56:135–147. Perry, K., Y. Hwang, F. D. Bushman, and G. D. Van Duyne. 2006. Structural basis for specificity in the poxvirus topoisomerase. Mol. Cell 23:343–354. Prescott, D. M., J. Kates, and J. B. Kirkpatrick. 1971. Replication of vaccinia virus DNA in enucleated L-cells. J. Mol. Biol. 59:505–508. Sakaguchi, A., M. Miyaike, K. Kuroda, N. Nozaki, M. Tanaka, M. Hibino, Y. Fujii, S. Kato, and A. Kikuchi. 2002. Epitope distribution of randomly established monoclonal antibodies against human type II DNA topoisomerases. J. Biochem. (Tokyo) 132:409–416. Sekiguchi, J., N. C. Seeman, and S. Shuman. 1996. Resolution of Holliday junctions by eukaryotic DNA topoisomerase I. Proc. Natl. Acad. Sci. USA 93:785–789. Sekiguchi, J., and S. Shuman. 1997. Domain structure of vaccinia DNA ligase. Nucleic Acids Res. 25:727–734. Shuman, S. 1995. Vaccinia virus DNA ligase: specificity, fidelity, and inhibition. Biochemistry 34:16138–16147. Shuman, S. 1998. Vaccinia virus DNA topoisomerase: a model eukaryotic type IB enzyme. Biochim. Biophys. Acta 1400:321–337. Smith, G. L., Y. S. Chan, and S. M. Kerr. 1989. Transcriptional mapping and nucleotide sequence of a vaccinia virus gene encoding a polypeptide with extensive homology to DNA ligases. Nucleic Acids Res. 17:9051–9062. Tsai, H. J., W. H. Huang, T. K. Li, Y. L. Tsai, K. J. Wu, S. F. Tseng, and S. C. Teng. 2006. Involvement of topoisomerase III in telomere-telomere recombination. J. Biol. Chem. 281:13717–13723. Welsch, S., L. Doglio, S. Schleich, and J. Krijnse Locker. 2003. The vaccinia virus I3L gene product is localized to a complex endoplasmic reticulumassociated structure that contains the viral parental DNA. J. Virol. 77:6014– 6028.
