Effective Formation of the Segregation-Competent Complex Determines Successful Partitioning of the Bovine Papillomavirus Genome during Cell Division

Abstract

Effective segregation of the bovine papillomavirus type 1 (BPV1), Epstein-Barr virus (EBV), and Kaposi's sarcoma-associated human herpesvirus type 8 (KSHV) genomes into daughter cells is mediated by a single viral protein that tethers viral genomes to host mitotic chromosomes. The linker proteins that mediate BPV1, EBV, and KSHV segregation are E2, LANA1, and EBNA1, respectively. The N-terminal transactivation domain of BPV1 E2 is responsible for chromatin attachment and subsequent viral genome segregation. Because E2 transcriptional activation and chromatin attachment functions are not mutually exclusive, we aimed to determine the requirement of these activities during segregation by analyzing chimeric E2 proteins. This approach allowed us to separate the two activities. Our data showed that attachment of the segregation protein to chromatin is not sufficient for proper segregation. Rather, formation of a segregation-competent complex which carries multiple copies of the segregation protein is required. Complementation studies of E2 functional domains indicated that chromatin attachment and transactivation functions must act in concert to ensure proper plasmid segregation. These data indicate that there are specific interactions between linker molecules and transcription factors/complexes that greatly increase segregation-competent complex formation. We also showed, using hybrid E2 molecules, that restored segregation function does not involve interactions with Brd4.

Several animal DNA viruses, including bovine papillomavirus type 1 (BPV1), Epstein-Barr virus (EBV), and Kaposi's sarcoma-associated human herpesvirus type 8 (KSHV), maintain their genomes as nuclear extrachromosomal multicopy plasmids in latently infected proliferating host cells. Each of these viruses carries a sequence-specific DNA binding protein which contains the domain required to bind specific receptor complexes on mitotic chromatin (reviewed in references 15, 27, and 43). Viral genomes are tethered to host mitotic chromosomes by segregation proteins at multimeric binding sites. During cell division, the viral genome is partitioned into daughter cells. Therefore, it is widely accepted that segregation protein tethering and multimeric binding to the viral genome are necessary and sufficient to impart a segregation/partitioning function to viral genomes in proliferating cells. However, the identity of the mitotic receptor and how it interacts with the segregation proteins remain elusive.

The E2 protein provides a segregation/partitioning function in BPV1 by binding of E2 multimeric binding sites. E2 is a multifunctional sequence-specific DNA binding protein that is required for viral genome replication (10, 62, 67-69), transcriptional activity (60, 65), and viral genome maintenance (2, 23, 38, 51, 59). The E2 protein is comprised of an N-terminal transactivation domain (TAD), a flexible hinge region, and a C-terminal DNA binding/dimerization domain (19). BPV1 E2 associates with mitotic chromosomes throughout mitosis, and this interaction is dependent on sequences within the TAD of E2 (2, 8, 23, 38, 59). E2 proteins from several types of human papillomavirus (HPV) interact with mitotic chromosomes (14, 46, 70); however, the pattern of mitotic chromosome binding for HPV E2 proteins is distinct from that for BPV1 E2, and the pattern varies among E2 proteins from different HPV genera (46, 55).

EBV and KSHV also carry single viral proteins that are responsible for viral genome maintenance—EBNA1 and LANA1, respectively (15). Like E2, both proteins are involved in viral DNA replication, transcriptional regulation, and viral genome maintenance. The viral genome segregation/partitioning function has been attributed directly to the ability of EBNA1 to associate with host cell chromosomes during mitosis (20, 50), and LANA1 provides the same function (5, 11).

Viral segregation protein association with host chromosomes is mediated via direct interaction with cellular factors. Several independent studies have focused on identifying cellular partners as potential receptors for these viral segregation proteins (reviewed in reference 15). It is remarkable that despite very similar segregation mechanisms in these viruses, identified cellular partners differ. Interestingly, it was recently shown that the cellular double-bromodomain protein Brd4 interacts with E2, EBNA1, and LANA1 proteins (39, 47, 54, 74). Brd4 is required for attachment of HPV31, HPV16, and BPV1 E2 to mitotic host DNA (1, 9, 74). Subsequent studies showed that E2 proteins from several papillomaviruses can interact with Brd4 and that this interaction is required for E2-mediated transcriptional regulation (24, 44, 54, 56, 73). However, Brd4 is not required for the maintenance of all HPV genomes (44). With EBNA1, silencing of Brd4 does not affect mitotic chromosome attachment, suggesting that Brd4 is not required for EBNA1 attachment to chromosomes (39). These data suggest that Brd4 plays an important role in the viral life cycle; however, the role of Brd4 and other proteins in viral segregation requires further investigation.

The E2 TAD is critical for E2-mediated DNA replication, transcriptional regulation, chromatin attachment, and segregation. E2 TAD mutation reveals the remarkable complexity of this domain. A single amino acid substitution leads to a loss of protein function (2, 9, 75). Interestingly, E2 seems to be more sensitive to point mutations with regard to chromatin attachment, partitioning, and transcription activation than replication initiation function (2). However, it is difficult to completely separate transactivation and chromatin attachment/segregation functions. For example, mutations in the N terminus of E2 that affect E2 transcriptional activity always impact partitioning (2), indicating that E2 chromatin attachment and transcriptional activity are not mutually exclusive, and both activities are required for E2-mediated segregation.

Our previous studies revealed that BPV1 E2 segregation/partitioning function can be eliminated by two different mutations (2). First, E2 point mutations that abolish binding to mitotic chromosomes disrupt segregation/partitioning function. Second, point mutations that eliminate transcriptional activation function while maintaining chromatin binding function can disrupt segregation/partitioning. This suggests that segregation/partitioning of viral genomes may depend on both chromatin receptors and transcriptional machinery.

The LANA1 chromatin attachment domain resides within 22 N-terminal amino acids (aa), and it has been shown to interact with the H2A-H2B dimer (6, 52). The herpes simplex virus 1 (HSV1) VP16 protein activation domain is a strong acidic transactivator that interacts with several transcriptional machinery components. We performed complementation studies to determine how E2 mutations affect chromatin attachment by using well-defined domains from heterologous proteins. We engineered chimeric E2 proteins in which the E2 TAD was replaced or complemented with the chromatin attachment domain from LANA1 of KSHV or with the TAD from the HSV1 protein VP16. The recombinant proteins allowed us to separate and analyze the requirements of each of these domains during segregation. Our data show that chromatin attachment is necessary but not sufficient to promote segregation protein activity and subsequent segregation/partitioning of the viral genome. We also demonstrate that assembly of the segregation complex carrying multimeric copies of the segregation protein is a necessary prerequisite to segregation/partitioning. Furthermore, segregation activity could be eliminated by mutation of several domains and complemented by heterologous domains with similar activities. We concluded that coordinated interactions between chromatin attachment and transcription complex interaction domains promote viral segregation protein function.

MATERIALS AND METHODS

Plasmids.

The backbone of the GTU plasmid vector system that was used for partitioning assays was described previously (42). Partitioning assay control plasmids were described by Abroi et al. (2). GTU plasmids used in this study have slight modifications in the locations of functional elements. The general structure and functional elements of GTU plasmids are represented in Fig. 4A, and a schematic representation of the designed chimeric E2 proteins is shown in Fig. 1. The chromatin interacting region of LANA1 (amino acids 1 to 22) was obtained by oligonucleotide annealing followed by cloning into the GTU-E2 plasmid after digestion with XbaI-Eco91I (E2 starts from amino acid 222). The resulting plasmid (GTU-LANA:E2C) was verified by sequencing. All transactivation domains used in this study (for VP16, the last 80 amino acids of the C-terminal part; for human p53, amino acids 1 to 58; and for human c-Myc, amino acids 1 to 167) were amplified by PCR and cloned into either the Ecor91I or XbaI-Eco91I sites in the GTU-LANA:E2C plasmid. Resulting plasmids did (Ecor91I) or did not (XbaI-Eco91I) contain LANA1 chromatin attachment sequences. The full-length E2-expressing GTU plasmids encoding N-terminal amino acid substitutions were constructed by replacing the Eco81I-XmaJI fragment from the wild-type (wt) E2 open reading frame in the GTU-E2 or GTU-LANA:E2 plasmid (containing the LANA chromatin attachment domain in the N terminus of the full-length wt E2 open reading frame) with the Eco81I-XmaJI fragment from the pRetE2 plasmid (2). The VP16 TAD was inserted into the hinge region of the full-length E2 open reading frame by being cloned into the Ecor91I site (the TAD of VP16 starts after amino acid 220). The phenylalanine-to-alanine substitution at position 442 in the VP16 TAD was created by PCR. The first intron of hEF1α was amplified by PCR and cloned into the XbaI site in the GTU vector. The frameshift mutation in the GTU coding sequence was created by Eco91I restriction followed by Klenow fragment filling and religation. The construct was verified by sequencing.

FIG. 1.

Schematic representation of chimeric E2 proteins. Numbers indicate positions in the amino acid sequence. Subscripted numbers and text indicate the locations of mutations with regard to amino acid sequence or the origin of the indicated TAD. TAD, transactivation domain; DBD, DNA binding domain; LANA_1-22, LANA1 chromatin attachment domain.

Mobility shift and immunoprecipitation assays were performed using pCG-based expression plasmids (64). Plasmids pCGE2, pCGE2C, pCGVP16:E2C, and pCGR37A were described previously (2, 36, 68, 69). The expression plasmid pCGLANA:E2C was constructed by replacing the XbaI-BshTI fragment in pCGE2 with the XbaI-BshTI fragment from GTU-LANA:E2C (E2 starts from amino acid 222). The expression plasmid pCGR37A:VP16 was constructed by replacing the XbaI-XagI fragment in pCGE2 with the XbaI-XagI fragment from GTU-R37A:VP16 (the TAD of VP16 starts after amino acid 220). All constructed plasmids were verified by sequencing.

Cell lines and transfection.

Jurkat and U2OS cells were grown in Iscove's modified Dulbecco's medium (IMDM). CHO cells were grown in Ham's F-12 medium. All media were supplemented with 10% fetal calf serum (FCS). Electroporation experiments were carried out as described previously (68), using a Bio-Rad Gene Pulser II apparatus supplied with a capacitance extender. Jurkat cells were transfected with a capacitance of 1,000 μF at 210 V. CHO cells were transfected with a capacitance of 975 μF at 230 V. U2OS cells were transfected with a capacitance of 975 μF at 180 V. All transfection mixtures contained 50 μg of denatured salmon sperm DNA.

For the transactivation assay, Jurkat cells were cotransfected with 25 ng of pRL-TK, 250 ng of E2-responsive reporter plasmid (contains five E2 binding sites in front of the simian virus 40 [SV40] early promoter and Luc gene), and 250 ng of GTU plasmid. Approximately 24 to 48 h later, cells were harvested and lysed by freeze-thawing, and luciferase activity was analyzed using a TD-20/20 luminometer (Turner Designs) according to the methods in Promega's dual-luciferase reporter assay manuals.

Mobility shift assay.

Mobility shift assays were performed as described previously (3, 36). Briefly, COS-7 cell extracts were transfected with 1 μg expression plasmid pCG and lysed in 100 μl of lysis buffer (50 mM Tris-HCl [pH 7.5], 100 mM KCl, 0.1 mM EDTA, 0.35% Nonidet P-40, 10 mM dithiothreitol [DTT], and protease inhibitors [Roche]) on ice for 30 min. For gel shift assays, 0.2 μl of cell extract was incubated in 10 μl binding buffer (10 mM Tris-HCl [pH 7.5], 100 mM KCl, 2 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride [PMSF], 15% glycerol, 5 mg of bovine serum albumin [BSA] per ml, protease inhibitors [Roche]) at room temperature for 20 min in the presence of 1 μg of sonicated salmon sperm DNA and 0.5 ng of a ³²P-labeled, double-stranded, high-affinity probe for BPV1 E2 binding site 9 (BS9; 5′-ACAAAGTACCGTTGCCGGTCGAA-3′ [binding site is shown in bold]). One microgram of anti-BPV1 E2 protein monoclonal antibody (MAb) 1E4 (36) was added to each reaction mix simultaneously with the cell extracts and probe. Protein-DNA complexes were resolved by 6% PAGE (acrylamide-N,N-methylene-bisacrylamide; 80:1) in 0.25× Tris-borate-EDTA buffer (TBE). Gels were dried and exposed to X-ray film.

Cooperative DNA binding assay.

Cell lysates for DNA binding assays were prepared from transfected COS-7 cells. Thirty-six hours later, cells were lysed in 50 μl cold lysis buffer (45 mM Tris-HCl, pH 8.0, 0.45 M NaCl, 0.9% Igepal, 10% glycerol, 5 mM EDTA, 0.5 mM PMSF, protease inhibitors [Roche]). The binding reaction was carried out for 15 min in 15 μl binding buffer (10 mM HEPES, pH 8.0, 5 mM DTT, 0.5 mM PMSF, 0.1 M NaCl, 10% glycerol, 0.1 ng/μl carrier DNA, protease inhibitors [Roche]). A radiolabeled probe containing two E2 binding sites (GATCTGTACCGGCAACGGTCGGATCTGTAACCGGCAACGGTCGGATC [binding sites are shown in bold]) was created using PCR. Gel electrophoresis was carried out by 6% PAGE (80:1) in 0.25× TBE. Gels were fixed with 10% ethanol-30% acetic acid, dried, visualized, and quantified using a Typhoon 9400 instrument (GE Healthcare) with product software. The maximal occupancy was determined, and k_ab was calculated according to the following formula: k_ab = [1/(maximum occupancy − 1)]² (45).

ChIF analysis.

Chromatin immunofluorescence (ChIF) analysis was performed as described previously (2). Briefly, CHO cells were transfected with 1 μg GTU plasmid. Approximately 18 h after transfection, cells were treated with colcemid (final concentration, 0.1 μg/ml) to enrich the fraction of mitotic cells. Cells were collected with phosphate-buffered saline (PBS)-EDTA (3 mM), pelleted by centrifugation, suspended in 5 ml of 0.075 M KCl solution, incubated for 15 min at room temperature, pelleted by centrifugation, resuspended in a small volume of 0.075 M KCl solution, and transferred to polylysine-coated slides. Cells were then incubated for 10 min at room temperature and fixed in −20°C cold methanol for 10 min. For immunostaining, cells were incubated with a mixture of primary antibodies 3E8 and 5H4 (5 ng/μl [each]) (36) in antibody-binding solution (2% BSA in PBS) for 1 h at room temperature and then washed with PBS. The cells were incubated with Alexa 488-conjugated secondary antibody (Invitrogen) in antibody-binding solution for 1 h at room temperature and then washed with PBS. Cells were then placed under coverslips with SlowFade Gold mounting medium with DAPI (4′,6-diamidino-2-phenylindole) (Invitrogen). Cells were analyzed using a Nikon Eclipse TE2000-U fluorescence microscope equipped with an appropriate filter set.

Immunofluorescence analysis of anaphase cells.

CHO cells were transfected with 1 μg of GTU plasmid. Following transfection, cells were plated on polylysine-coated slides (placed into a 100-mm cell culture dish). At approximately 18 h posttransfection, cells were treated with nocodazole (final concentration, 50 ng/ml) to increase the fraction of mitotic cells. Five hours after nocodazole treatment, cells were washed three times with PBS to remove nocodazole and to release cells from the cell cycle block. Thirty minutes after nocodazole removal, cells were fixed in 4% paraformaldehyde. Prior to immunostaining, cells were permeabilized with 0.5% Triton X-100 in PBS. Fixed cells were incubated with monoclonal anti-α-tubulin antibody (clone B-5-1-2; Sigma) diluted 1:2,500 and detected with Alexa 488-conjugated secondary antibody (Invitrogen). Cells were then washed three times (5 min/wash) with PBS. For E2 or E2 derivatives, a mixture of Alexa 568-conjugated 3E8 and 5H4 antibodies (5 ng/μl [each]) (36) was used. Primary 3E8 and 5H4 antibodies were labeled using an Alexa Fluor 568 monoclonal antibody labeling kit (Invitrogen). All antibodies were diluted in antibody-binding solution (2% BSA in PBS). Cells were placed under coverslips with SlowFade Gold mounting medium with DAPI (Invitrogen). Cells were analyzed using a Nikon Eclipse TE2000-U fluorescence microscope equipped with an appropriate filter set.

FISH.

CHO cells were transfected with 1 μg of GTU plasmid. Approximately 18 h after transfection, cells were treated with colcemid (final concentration, 0.1 μg/ml) to enrich the fraction of mitotic cells. Prior to fixation, cells were treated as described for ChIF. Cells were fixed in −20°C methanol-glacial acetic acid (3:1 [vol/vol]). Chromosome spreads were prepared by dropping the cell suspension on wet polylysine-coated slides, followed by quick drying on a hot metal plate. Hybridization probes were generated by nick translation, using biotin-16-dUTP as a label and GTU plasmid as a template. The final probe fragment size was adjusted to 200 to 500 bp by DNase I digestion. Fixed cells were incubated for 20 min in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) at 60°C, followed by dehydration in a 70%, 80%, and 96% ethanol series (−20°C) for 1 min at each concentration. Slides were then air dried. Dehydrated cells were treated with RNase A (100 μg/ml) for 1 h at 37°C in 2× SSC, washed three times for 5 min each with 2× SSC at 37°C, dehydrated in ethanol as described above, treated with pepsin (50 μg/ml in 0.01 M HCl) for 8 min at 37°C, washed for 5 min with PBS, treated with 0.95% formaldehyde for 10 min in PBS, washed for 10 min with PBS, dehydrated by an ethanol series (−20°C; 2 min at each ethanol concentration), and air dried. Finally, prior to hybridization, chromosome preparations were denatured at 75°C in 70% formamide-2× SSC for 3 min and immediately dehydrated with an ethanol series (2 min at each concentration) at room temperature. After denaturation, 20 μl DNA probe mix (composed of 50% formamide in 2× SSC, 100 ng denatured DNA probe, and 5 μg denatured herring sperm carrier DNA) was applied to each slide, sealed under a coverslip, and hybridized overnight at 37°C in a moist chamber. Subsequent fluorescence in situ hybridization (FISH) procedures were performed according to the manufacturer's protocol for a TSA 22 kit (Invitrogen). Cells were placed under coverslips with SlowFade Gold mounting medium with DAPI (Invitrogen), and cells were analyzed with a Nikon Eclipse TE2000-U fluorescence microscope equipped with an appropriate filter set.

Plasmid partitioning assay.

Partitioning assays were performed as described previously (2). Briefly, Jurkat cells were transfected by electroporation with 1 μg of GTU plasmid (3 × 10⁶ cells per transfection). At set time points after transfection, equal aliquots of the cell suspension from all transfected samples were collected for flow cytometry analysis, and remaining cells were diluted with fresh medium, maintaining the cell concentration at 0.3 × 10⁶ to 1 × 10⁶ cells per ml. The number of enhanced green fluorescent protein (EGFP)-positive cells per milliliter was calculated for each sample according to the following formula: [(number of positive cells × 1,000)/(time × flow rate)] × dilution (time is measured in minutes, and flow is measured in microliters/minute). The numbers of positive cells at different time points were normalized to that for the first time point posttransfection. Cell doubling rates were calculated by analyzing the increase in cell number, with the first time point (∼20 h after transfection) used as a starting point.

Immuno-FISH analysis of anaphase cells and colonies.

CHO cells were transfected with 1 μg GTU plasmid. To enrich the fraction of cells in anaphase, nocodazole treatment was used as described above. For colony analysis, destabilized EGFP (d1EGFP)-positive cells were sorted using a cell sorter (FACS Aria; BD Biosciences) and cultured on polylysine-coated slides (placed into a 100-mm cell culture dish) at low densities to ensure single-cell analysis. Cells were grown to colonies (6 to 8 cells) and fixed with methanol at −20°C for 10 min. E2 proteins were immunostained by fixing and blocking cells in 3% BSA-PBS for 30 min at room temperature, followed by incubation in a mixture of Alexa 568-conjugated 3E8 and 5H4 antibodies (5 ng/μl [each]) (36) in 3% BSA-PBS for 1 h at room temperature. Cells were then washed three times for 5 min each with PBS at room temperature, fixed with −20°C cold methanol-glacial acetic acid (3:1 [vol/vol]) for 10 min and with 4% paraformaldehyde at room temperature for 10 min, followed by dehydration in a 70%, 80%, and 96% ethanol series (−20°C) for 2 min at each concentration, and air dried. FISH reactions were then performed as described above. Finally, cells were analyzed with a Nikon Eclipse TE2000-U fluorescence microscope equipped with an appropriate filter set.

Immunoprecipitation.

U2OS cells were transfected with 2 μg of the expression plasmid pCG. Cells on a 100-mm plate were washed twice with PBS. Cells were then collected with PBS-EDTA and lysed in ice-cold lysis buffer (20 mM Tris-HCl [pH 8.0], 10% glycerol, 5 mM MgCl₂, 0.1% Tween 20, 0.1 M KCl, 0.2 mM PMSF, 0.5 mM DTT, and 1× Complete EDTA-free protease inhibitor mix [Roche]). Samples were sonicated on ice three times for 5 s each and centrifuged for 10 min at 14,000 × g at 4°C. Four micrograms of a mixture of Brd4-specific rabbit polyclonal primary antibodies, including Brd4B (specific to Brd4 peptide amino acids 225 to 347; LabAS, Estonia), Brd4C (specific to Brd4 peptide amino acids 1359 to 1400; LabAS, Estonia), and AB46199 (Abcam), was added to the supernatant. The cells were incubated for 3 h at 4°C with gentle rocking. Ten microliters of protein Sepharose G beads (GE Healthcare) was added to the sample and incubated for an additional 3 h at 4°C. The beads were washed three times with ice-cold lysis buffer and resuspended in 1× SDS sample buffer. The samples were heated at 100°C for 10 min, and proteins were separated in a 10% SDS-PAGE gel. Separated proteins were transferred to an Immobilon-P membrane (Millipore). E2 and its derivatives were detected with 1E4 antibody (36).

Large-scale high-content image analysis.

All samples were prepared as described for ChIF. E2 and its derivatives were detected by the DNA binding domain (DBD)-specific antibodies 3E8 and 5H4 (36) and were visualized by an Alexa 568-conjugated secondary antibody. Cellular DNA was labeled with DAPI. ChIF slides were imaged using a Thermo Scientific Cellomics ArrayScan HCS reader and were analyzed using the SpotDetector.V3 algorithm. For each sample, at least 720 pictures were taken at a magnification of ×40, with autofocusing between every five pictures. Mitotic cells (“objects” in Protocol S1 in the supplemental material) in DAPI channel pictures and E2- or E2 derivative-specific dots in Alexa 568 channel pictures were identified and analyzed automatically by Cellomics ArrayScan software according to the assay protocol parameters (see Protocol S1 in the supplemental material). Pictures with false object identification were excluded from analysis. The detailed ArrayScan HCS reader protocol is available as Protocol S1 in the supplemental material.

Immunoblotting.

Total proteins from equal numbers of cells were separated by 10% SDS-PAGE and transferred to an Immobilon-P membrane (Millipore). Western blotting was performed using monoclonal 1E4 (36) and β-actin (Sigma-Aldrich) antibodies. Peroxidase-conjugated secondary antibody (LabAS Ltd., Estonia) signals were detected using an ECL detection kit (GE Healthcare).

RESULTS

Replacement of the E2 TAD with the LANA1 chromatin binding domain does not interfere with the ability of the protein to bind specific DNA sequences.

The role of E2 in viral genome maintenance is to act as a linker molecule that tethers the viral genome to host chromosomes. The E2 TAD is responsible for chromatin attachment and C-terminal DNA binding, while the dimerization domain is responsible for binding the viral genome at E2 binding sites (38, 59). Similarly, LANA1 (from KSHV) serves as a linker protein between the viral genome and host mitotic chromosomes. The first 22 N-terminal amino acids of LANA1 are sufficient for chromatin attachment (6, 7, 52). Thus, we reasoned that replacement of the E2 TAD with the LANA1 N-terminal 22 amino acids should retain chromatin attachment capabilities and that the resulting LANA:E2C chimeric protein should functionally replace E2 during segregation/partitioning.

The chromatin attachment domain from LANA1 was fused to the E2C protein, and the sequence-specific DNA binding activity of LANA:E2C was assayed using gel shift analysis (Fig. 2 A and B). LANA:E2C effectively bound the radiolabeled BS9 probe and caused a band shift, indicating that LANA:E2C was able to bind E2 binding sites (Fig. 2A, lane 5). A similar band shift pattern was observed when wt E2 and E2C were used in place of LANA:E2C (Fig. 2A, lanes 3 and 4). Lysates from mock-transfected cells and probe alone were used as negative controls (Fig. 2A, lanes 1 and 2). Addition of the monoclonal antibody 1E4 (36), which recognizes an epitope in the hinge region of E2, supershifted the E2-BS9 complex (Fig. 2B, lanes 3 to 5). These data demonstrate that fusion of LANA1 N-terminal amino acids 1 to 22 to E2C did not interfere with protein folding or E2 binding.

FIG. 2.

Chimeric E2 proteins are able to bind DNA. (A) DNA binding assay using chimeric E2 proteins. The sequence-specific binding of E2 and its derivatives to one palindromic DNA target was determined by mobility shift assays. E2 proteins were expressed in COS-7 cells. Band shift assays were performed with 0.2 μl of cell extract and 0.5 ng of radiolabeled BS9 (5′-ACAAAGTACCGTTGCCGGTCGAA-3′) for 20 min at room temperature. (B) Reactivity of MAb 1E4 (36) with hybrid E2 proteins expressed in COS-7 cells. Hybrid E2 proteins were mixed with relevant DNA targets, and MAb 1E4 was added simultaneously. Oligo, oligonucleotide; mock, cells which were transfected with carrier DNA only. Protein-DNA complexes were resolved by 6% PAGE in 0.25× Tris-borate-EDTA. (C) Measurement of cooperative DNA binding of E2, E2C, and LANA:E2C at two neighboring binding sites. Graphs represent quantified data from cooperative DNA binding assays performed with different amounts of E2, E2C, and LANA:E2C COS-7 lysates (0.0625 to 2 μl) and with two E2-binding-site-radiolabeled probes. Open triangles, free probe; filled triangles, one E2 binding site occupied; filled squares, two E2 binding sites occupied. The maximal occupancy of the single E2 binding site and the respective k_ab were calculated according to the method of Monini et al. (45).

E2 binding to multimeric binding sites shows clear cooperativity (45). In addition, we studied the cooperativity of E2, E2C, and LANA:E2C proteins for binding to dimeric sites. As shown in Fig. 2C, the E2 protein was capable of binding to an oligonucleotide carrying two E2 sites, with clear cooperativity, while E2C and LANA:E2C bound both sites with significantly reduced cooperativity (Fig. 2C).

Chimeric LANA:E2C protein attaches to host chromosomes and facilitates plasmid-chromosome attachment in an E2-binding-site-dependent fashion.

The BPV1 E2 protein is capable of binding to host mitotic chromatin. It is thought that this activity is sufficient for E2-mediated viral genome tethering to host mitotic chromosomes (23, 38, 59). We asked whether LANA:E2C attached to host chromosomes and whether or not it could facilitate plasmid tethering to mitotic chromatin. To evaluate the ability of LANA:E2C to bind mitotic chromosomes, ChIF analysis was performed (Fig. 3, upper panels). As shown in Fig. 3, LANA:E2C molecules appeared as green dots (Alexa Fluor 488 stain) on DAPI-stained mitotic chromosomes. The LANA:E2C protein was localized on metaphase chromosomes, similar to BPV1 E2, which was used as a positive control (Fig. 3, upper panels). The E2C protein (Fig. 3, upper panels) failed to bind chromatin. These data suggest that the LANA1 domain effectively replaces the E2 TAD to provide chromatin binding function to the chimeric protein. The ability of LANA:E2C to facilitate plasmid attachment was analyzed by labeling metaphase spreads with biotin-labeled plasmid-specific FISH probes (Fig. 3, lower panels). FISH signals were amplified using the tyramide enhancement FISH method. Discrete green dots (the Alexa Fluor 488 label) corresponding to the specific signal from the LANA:E2C expression plasmid were clearly detected on metaphase chromosomes (Fig. 3, lower panels). Similar staining was observed in E2 positive controls (Fig. 3, lower panels). The E2C protein completely failed to mediate plasmid attachment to chromosomes (Fig. 3, lower panels). These data indicate that the heterologous chromatin attachment domain in the LANA1 N terminus effectively replaces the E2 TAD for chromatin attachment and tethering of E2-binding-site-containing plasmids to metaphase chromosomes.

FIG. 3.

Chimeric LANA:E2C protein attaches to host chromosomes and supports plasmid-chromatin attachment. CHO cells were transfected with 1 μg of partitioning assay GTU plasmid expressing BPV1 E2C, BPV1 wt E2, or chimeric LANA:E2C protein. Forty-eight hours after transfection, cells were treated with colcemid and analyzed by ChIF (upper panels) and FISH (lower panels) for the presence of proteins and plasmid DNA, respectively. ChIF and FISH signals are visible as green dots (Alexa Fluor 488 stain). Chromosomes were counterstained with DAPI, and slides were analyzed using a Nikon Eclipse TE2000-U confocal fluorescence microscope equipped with an appropriate filter set.

LANA:E2C functionality is disturbed for plasmid segregation function.

Chromatin attachment of BPV1 genomes to host DNA is absolutely required for proper viral genome segregation/partitioning, and it is believed to be sufficient for this biological function. We measured the segregation/partitioning function of LANA:E2C. We developed an assay that allowed us to measure segregation/partitioning of the plasmid in dividing cell populations (2). Jurkat cells were transfected with nonreplicating plasmid GTU vectors that were described previously (2). The GTU vector contains three functional elements that are important for this particular assay: (i) the cytomegalovirus (CMV) promoter to control expression of d1EGFP; (ii) a Rous sarcoma virus long terminal repeat (RSV LTR)-driven expression unit for E2 or derivatives, such as LANA:E2C; and (iii) an array of E2 binding sites introduced into the plasmid (Fig. 4 A). When E2 is expressed from GTU, it simultaneously binds to its binding sites on the GTU plasmid and chromosomes. These interactions ensure proper segregation of the GTU plasmid and enable it to travel to daughter cells during cell division. Because only the cells that contain the GTU plasmid exhibit the short-living d1EGFP fluorescence, segregation/partitioning function is measured as the change in the number of d1EGFP-positive cells during cell division. Our data, which demonstrated binding of the LANA:E2C protein and tethering of the E2-binding-site-containing plasmid to the chromosomes, suggested that LANA:E2C should be fully functional for segregation.

FIG. 4.

The chimeric LANA:E2C protein does not support plasmid partitioning in functional assays. (A) Map of the GTU plasmid used for partitioning assays. CMV, CMV immediate-early promoter; d1EGFP, destabilized EGFP coding sequence; intron, rabbit β-globin IV intron; pA, thymidine kinase poly(A) signal; RSV LTR, Rous sarcoma virus 5′-long-terminal-repeat promoter; KanaR, kanamycin resistance gene; 10E2BS, oligomerized E2BS9 (10 copies). (B) Plasmid partitioning assay. Partitioning assays were performed as described previously (2). Jurkat cells were transfected with 1 μg GTU plasmid. The total cell number and number of EGFP-positive cells per milliliter were determined by fluorescence-activated cell sorting analysis at different time points. From these data, the numbers of cell doublings and EGFP-positive cells were calculated relative to the first values (∼20 h posttransfection). The number of EGFP-positive cells was plotted against the number of cell doublings. Plasmids expressing E2 and E2C were used as positive and negative controls, respectively. (C) Colony immuno-FISH analysis. Cells were transfected with GTU plasmid. GFP-positive cells were sorted by FACS. Single cells were cultured on slides and expanded to colonies, followed by combined IF-FISH analysis of E2 proteins and plasmid DNA. Localization of E2 proteins is represented by the red signal (Alexa Fluor 568; second column). Localization of plasmid DNA is represented as the green signal (Alexa Fluor 488; third column). DNA was counterstained with DAPI (first column). Merged images are in the fourth column. E2- and E2C-expressing plasmids were used as positive and negative controls, respectively. (D) Western blot analysis of partitioning assay samples. Western blot samples were collected from Jurkat cells ∼17 h after transfection. Bars on the right indicate the locations of marker bands (kDa).

The number of d1EGFP-positive cells was measured by fluorescence-active cell sorter (FACS) analysis at different time points after transfection. As shown in Fig. 4B, wt E2 plasmids were effectively segregated into daughter cells. Thus, the number of d1EGFP-positive cells increased for at least six doublings. On the other hand, the N-terminally truncated reporter plasmid, which is defective for plasmid-chromatin attachment, failed to segregate into daughter cells, and the number of d1EGFP-positive cells decreased after the second cell doubling (Fig. 4B). These data suggest that the reporter plasmids expressing E2 and E2C are valid positive and negative controls for segregation/partitioning function estimation, respectively. To our surprise, we saw that LANA:E2C behaved identically to the negative-control E2C protein, despite the fact that attachment to chromosomes and the ability to mediate plasmid-mitotic DNA association were normal (Fig. 4B). Western blot analysis showed that the LANA:E2C expression level was significantly higher than that of E2. E2 expression was typically undetectable when expressed from the RSV LTR promoter in Jurkat cells (Fig. 4D).

To ensure that the segregation/partitioning assay reflects plasmid segregation, we decided to monitor plasmids and segregation proteins directly in the cells via colony immuno-FISH analysis (Fig. 4C). We transfected CHO cells with E2, E2C, or LANA:E2C GTU plasmid. Twenty-four hours later, GFP-positive cells were isolated via FACS. Cells were plated at low densities on microscope slides. Small colonies formed from single cells after several rounds of cell division over a 3- to 4-day period. The cells were fixed with methanol and analyzed by immuno-FISH, which allowed for simultaneous detection of protein and plasmid DNA in the same cell. Hybridization of fixed cells with a GTU plasmid-specific probe showed a normal plasmid distribution to the daughter cells in the presence of E2 (Fig. 4C, uppermost panels). In contrast, plasmids that expressed E2C or LANA:E2C were not detected in peripheral colony cells (Fig. 4C, lowest and middle panels, respectively). Simultaneous analysis of protein localization by IF showed that E2 colocalized with plasmid DNA (Fig. 4C, first column). In a few colonies, we detected LANA:E2C protein in daughter cells; however, those cells did not contain the plasmid signal at a detectable level (data not shown). These immuno-FISH data confirm our previous results obtained from the segregation/partitioning assay of dividing Jurkat cells.

We showed that the LANA:E2C protein was bound to metaphase chromosomes and facilitated E2-binding-site-dependent plasmid-chromosome attachment. We speculated that the failure to facilitate plasmid segregation may have occurred during later phases of mitosis. The ChIF and FISH analyses (Fig. 3) were performed using cells that were blocked in prometaphase by colcemid. To test protein and plasmid compartmentalization in later phases of mitosis, transfected cells were blocked with nocodazole for up to 5 h. The cells were then released by washes with nocodazole-free PBS and medium. Cells were then fixed with either 4% paraformaldehyde (for IF analysis) or methanol (for immuno-FISH analysis). IF analysis showed that the chimeric LANA:E2C protein was still attached to host DNA in anaphase (Fig. 5 A). The immuno-FISH analysis indicated that LANA:E2C also supported plasmid DNA tethering to the host chromosomes in anaphase (Fig. 5B). These data suggest that the segregation/partitioning process is defective during late mitosis (likely during telophase) in LANA:E2C-expressing cells under the conditions used for this study.

FIG. 5.

IF and immuno-FISH analyses of anaphase cells. (A) Chimeric LANA:E2C protein attached to chromosomes. CHO cells were transfected with 1 μg of GTU partitioning assay plasmid expressing LANA:E2C, E2, or E2C. Eighteen hours after transfection, cells were treated with nocodazole for up to 5 h. Nocodazole was removed by PBS washes. Thirty minutes later, cells were fixed in 4% paraformaldehyde. Immunostaining was performed for E2 proteins (red; Alexa Fluor 568) (first column) and tubulin (green; Alexa Fluor 488) (second column). DNA was counterstained with DAPI (third column). Merged images are shown in the fourth column. E2- and E2C-expressing plasmids were used as positive and negative controls, respectively. (B) Chimeric E2 proteins support plasmid tethering to host chromosomes. Nocodazole treatment was performed as described above, except that cells were fixed with methanol. Localization of E2 proteins is represented as a red signal (Alexa Fluor 568; first column). Localization of plasmid DNA is represented as a green signal (Alexa Fluor 488; second column). DNA was counterstained with DAPI (third column). Merged images are shown in the fourth column. E2- and E2C-expressing plasmids were used as positive and negative controls, respectively.

Based on these results, we concluded that the wt E2 segregation complex was qualitatively different in composition from the LANA:E2C complex in the assays used here. While segregation/partitioning of plasmids into daughter cells was normal in the context of E2, the LANA:E2C segregation complex was unable to support distribution of plasmid DNA.

Segregation-competent complex formation and effectiveness differ between E2 and LANA:E2C.

BPV1 or heterologous replicon-based plasmids require at least five or six E2 binding sites to achieve proper E2-dependent segregation (51, 58). While E2 functions as a dimer, the presumable segregation-competent complex is composed of at least 10 to 12 E2 molecules.

We studied chromosome-bound complexes by using an automated fluorescence microscopic imaging system that enabled us to analyze hundreds of images and to measure both the number and intensity of protein complexes on mitotic chromosomes. We first analyzed E2 and LANA:E2C complexes. CHO cells were transfected with various concentrations of GTU plasmids and plated on ChIF slides. Images were obtained via high-content automated picture scans. Automated identification of mitotic cells was performed, and the number and intensity of E2 DBD-specific signals on mitotic chromosomes were scored (Fig. 6 A). At least 720 pictures of more than 4,000 mitotic cells were analyzed with each scan. Spot counts revealed that there were equal numbers of E2 and LANA:E2C spots in each mitotic cell (Fig. 6B), indicating that E2 and LANA:E2C have approximately equal numbers of attachment sites on chromosomes. The fact that E2 was expressed at much lower levels than LANA:E2C indicates that E2 has a much higher affinity for chromosome binding than LANA:E2C. Spot intensity measurements showed that E2 dots were notably brighter than LANA:E2C dots at low DNA concentrations (Fig. 6C, 0.5-μg series). However, at high DNA concentrations, dot intensities were similar (Fig. 6C, 1- and 2-μg series). Because both proteins have identical DBD epitopes, these results indicate that there are clear qualitative differences between E2 and LANA:E2C complexes on mitotic chromosomes. These data suggest that both E2 and LANA:E2C proteins are able to bind chromosomes and suggest that segregation-competent complex assembly may be concentration dependent. The E2 segregation-competent complex is functional at very low E2 levels, indicating that the E2 complex forms with a high affinity. In contrast, LANA:E2C must be present at high levels for segregation complex assembly to occur. The differences in E2 and LANA:E2C affinities must be considerable, since LANA:E2C expression was significantly higher than that of E2 (Fig. 4D).

FIG. 6.

High-content image analysis. (A) Examples of captured pictures and identification of mitotic cells and spots on mitotic chromosomes by an automated fluorescence microscope. ChIF slides were prepared, and E2 and its derivatives were detected with the DBD-specific antibodies 3E8 and 5H4 (36) and then visualized by Alexa 568-conjugated secondary antibody. Pictures were captured and analyzed automatically by a Thermo Scientific Cellomics ArrayScan HCS reader and an image analysis module. Image analysis included the following steps: (i) autofocusing and picture capture on the DAPI channel (first picture from left), (ii) picture capture on the Alexa 568 channel (second picture from left), (iii) identification of mitotic cells (third picture from left), and (iv) identification of spots and signal calculation (fourth picture from left). Blue lines in the DAPI channel picture indicate identified objects (mitotic cells). Orange lines show rejected objects. The red-lined area with red spots in the Alexa 568 channel picture shows the localization of identified and analyzed dots. The green line indicates the identification of more than 300 dots. (B) Quantification of average spot count per identified mitotic cell. (C) Quantification of total intensity per spot. CHO cells were transfected with 0.5, 1, and 2 μg of the appropriate GTU plasmids. ChIF slides were prepared for each transfection. All slides were analyzed identically by the automated fluorescence microscope as described above. Analyzed pictures from different samples were captured at a fixed exposure time and analyzed by an identical assay protocol (see Protocol S1 in the supplemental material). The standard deviation was calculated for 3 or 4 individually scanned fields.

To test this hypothesis, we modulated LANA:E2C expression by altering the LANA:E2C expression cassette. The first intron of human elongation factor 1α (hEF1α) was added to the 5′ end of the LANA:E2C coding sequence (iLANA:E2C). Modification of the expression cassette significantly increased LANA:E2C expression (Fig. 7 A, compare lanes 4 and 5). High-content image analysis showed that if LANA:E2C was expressed from the 5′-intron expression cassette, spot intensities were similar to those for E2 at low DNA concentrations (Fig. 6C, 0.5-μg series). As shown in Fig. 7B, iLANA:E2C supported GTU plasmid segregation almost as efficiently as E2. To ensure that this effect was due to increased protein levels and not to cis elements in the intron, several control plasmids were constructed. Addition of an hEF1α intron to E2C did not change its functionality (Fig. 7B), though increased E2C expression was observed (Fig. 7A). Expression of iLANA:E2C from the weaker thymidine kinase (TK) promoter did not increase plasmid segregation (Fig. 7B), indicating that iLANA:E2C-mediated segregation is dependent on protein levels. Finally, expression of iLANA:E2C containing a frameshift (FS) mutation did not increase plasmid segregation, indicating that full-length LANA:E2C is required for segregation (Fig. 7B).

FIG. 7.

High expression of LANA:E2C restores segregation. (A) The hEF1α intron increases LANA:E2C expression levels. At 48 hours posttransfection, samples from parallel partitioning assays were lysed for Western blot analysis. The arrow indicates the localization of actin, which was used as a loading control. (B) Partitioning assay of GTU plasmids containing hEF1α intron sequence.

Together, these data show that chromatin attachment alone does not ensure plasmid segregation. Rather, formation of a segregation-competent complex is required for segregation, and competent complex formation occurs at concentrations above a certain threshold. Wild-type E2 complex assembly is very efficient at low protein concentrations, and the E2 TAD is necessary for efficient complex assembly.

A functional E2 TAD is required for chromatin attachment and plasmid segregation.

Single amino acid substitutions in the N-terminal domain of E2 can specifically inactivate certain functions of the E2 protein (2); however, E2 transactivation and segregation/partitioning functions are difficult to separate. Mutants that affect transactivation always have a defect in partitioning function (2). Thus, we wondered if transcription activation is required for effective segregation/partitioning complex formation and function. To test this hypothesis, we complemented the defective E2 TAD for two functions—either transcriptional activation or chromatin attachment.

Most E2 N-terminal mutants that support BPV1 upstream regulatory unit (URR) replication are observed to be defective in transcriptional activation, chromatin attachment, plasmid chromatin attachment, or segregation/partitioning (2). Two mutants, the E39A and R68A mutants, are defective in chromatin attachment, plasmid chromatin attachment, and plasmid segregation/partitioning; however, they still maintain significant transcriptional activity (2). To restore chromatin attachment function to these mutants, we fused the chromatin attachment domain from LANA1 to the N termini of the two mutant constructs. This modification restored mitotic chromosome binding in both mutants (data not shown). Moreover, the segregation/partitioning assay demonstrated that segregation/partitioning function was partially restored (Fig. 8 A). The E39A and R68A mutants that lacked the chromatin attachment domain did not display proper segregation/partitioning (Fig. 8A). E39A and R68A mutant transcriptional activity was modified by addition of the VP16 transactivation domain (80 C-terminal amino acids) from HSV1 to the hinge region of the protein. Addition of the VP16 TAD had no effect on segregation/partitioning (Fig. 8A).

FIG. 8.

Partitioning assay of plasmids expressing chimeric E2 proteins. (A) Chromatin attachment and transactivation activities must act in concert to ensure plasmid segregation. The LANA chromatin attachment sequence was added to the N termini of point-mutated R37A, E39A, and R68A E2 proteins. The VP16 TAD was added to the hinge region of the point-mutated R37A and E39A E2 proteins. All mutants were characterized previously (2). To determine the partitioning properties of chimeric E2 proteins, partitioning assays were performed. (B) The p53 TAD or c-Myc TAD does not restore plasmid segregation function. The human p53 TAD (aa 1 to 57) and the c-Myc TAD (aa 1 to 167) were linked to E2C or LANA:E2C and tested for the ability to support plasmid segregation in partitioning assays. (C) The VP16 TAD does not interact with Brd4 and does not restore Brd4 interactions with E2 R37A mutant. U2OS cells were cotransfected with 3 μg of Brd4 expression plasmid and 3 μg of E2 or E2 derivative expression plasmid. Immunoprecipitation (IP) was performed using a rabbit polyclonal anti-Brd4 antibody. Western blot analyses of input (lanes 1 to 5) (left) and immunoprecipitated (lanes 6 to 10) (right) material were performed using the E2-specific antibody 1E4 (3). Arrows indicate the localization of marker bands (kDa).

In contrast to the E39A and R68A mutants, the R37A mutant does not display transcriptional activation but maintains at least 50% of the wild-type chromatin attachment activity. Addition of the LANA chromatin attachment domain to the N terminus of the R37A mutant had no effect on its ability to support plasmid segregation in partitioning assays (Fig. 8A), while addition of the VP16 TAD to the hinge region of this mutant completely restored partitioning/segregation function (Fig. 8A). These results suggest that both chromatin attachment and transactivation activities and the ability of E2 to interact with the transcription complex are necessary for effective segregation/partitioning function. Inactivation of either activity abolishes the formation of segregation-competent complexes and the ability to support segregation function.

High-level LANA:E2C expression is required for proper segregation activity in E2-binding-site-containing plasmids. As described above, addition of the functional VP16 TAD to the segregation-negative E2 R37A mutant completely restored R37A mutant partitioning function. Thus, we tested whether addition of the VP16 TAD to the hinge region of LANA:E2C restores LANA:E2C segregation function. The chimeric LANA:VP16:E2C protein is a strong transcriptional activator, displaying >100-fold higher transcription activation than that of LANA:E2C (Fig. 9 F).

FIG. 9.

Intact VP16 transactivation domain restores plasmid segregation. (A) LANA:VP16:E2C protein localization on host chromosomes in anaphase. Localization of LANA:VP16:E2C proteins is represented as a red signal (Alexa Fluor 568; first column). Tubulin is represented in green (Alexa Fluor 488; second column). (B) LANA:VP16:E2C protein tethers plasmid DNA to the chromosomes. Localization of LANA:VP16:E2C is represented as a red signal (Alexa Fluor 568; first column). Localization of plasmid DNA is represented as a green signal (Alexa Fluor 488; second column). (C and D) ChIF analysis of VP16 and LANA:F442A:E2C proteins. Localization of VP16 (C) and LANA:F442A:E2C (D) proteins is represented as a green signal (Alexa Fluor 488; first column). All immunoassays were performed in CHO cells transfected with 1 μg of GTU plasmid. DNA was counterstained with DAPI in all cases. (E) Partitioning properties of plasmids expressing E2 or its chimeric forms. The number of EGFP-positive cells per milliliter (y axis) was plotted against the number of cell doublings (x axis). (F) Transcription activation assay of E2 and its chimeric forms. Jurkat cells were transiently transfected with GTU plasmids and the E2-responsive reporter plasmid 5E2BS-Luc. Luciferase activity was measured 48 h later and normalized to Renilla luciferase expression from the non-E2-responsive thymidine kinase promoter of the cotransfected plasmid pRL-TK. The data are presented as promoter activities relative to the activity of the reporter alone.

IF analysis showed that the LANA:VP16:E2C protein was attached to host DNA during anaphase (Fig. 9A). Immuno-FISH analysis demonstrated that LANA:VP16:E2C also supported plasmid DNA tethering to host chromosomes (Fig. 9B). Thus, both chimeric LANA:E2C and LANA:VP16:E2C proteins efficiently mediate protein chromatin attachment and plasmid tethering. As shown in Fig. 9E, LANA:VP16:E2C partially restored GTU plasmid segregation, while LANA:E2C was completely inactive in the same assay. We engineered an additional chimeric protein, VP16:E2C, which is a very strong E2-binding-site-dependent transcriptional activator (Fig. 9F), but it was defective in chromatin attachment (Fig. 9C). Failure of VP16:E2C to support plasmid segregation (Fig. 9E) indicated that the transactivation activity/domain alone is not sufficient for segregation function.

The VP16 TAD contains a critical phenylalanine residue at position 442, and mutation of this amino acid to alanine reveals a good correlation between transactivation activity in vivo and binding of TFIID in vitro (13, 25). To test the effect of this mutation on segregation function, we generated the same mutation in LANA:VP16:E2C. Mutant LANA:F442A:E2C was approximately 4-fold less functional in transcription activation (from the E2-dependent promoter) than the wt protein (Fig. 9F). As expected, chromatin attachment of this mutant protein and the capability to mediate the association of plasmid with chromosomes were similar to those of E2 and other LANA chimeras (Fig. 9D). However, LANA:F442A:E2C was unable to support plasmid segregation function (Fig. 9E), indicating that the mutation eliminated interactions with cellular proteins that are required for plasmid partitioning.

Transactivation domains of human p53 and c-Myc do not restore plasmid segregation.

Transcriptional activation can be ensured through several different mechanisms. The herpesvirus VP16 TAD effectively restored segregation function to chromatin attachment-positive proteins. The VP16 TAD has an acidic domain that promotes a strong transcription activation capacity in hybrid proteins when it is attached to a heterologous DBD (12, 53). We constructed chimeric E2 proteins that carried TADs from other transcription-activating proteins. We then determined whether these domains could serve as functional plasmid segregation substitutes for the VP16 TAD. The TADs from p53 and c-Myc are well characterized. The p53 TAD has been mapped to amino-terminal residues 1 to 42 (66). Residues 1 to 143 of c-Myc are sufficient to provide transactivation activity in neoplastic transformation (28); however, additional studies extended the TAD of c-Myc to residues 1 to 167 (17). We added the TADs from p53 (aa 1 to 57) and c-Myc (aa 1 to 167) to the N terminus of E2C or to the hinge region of chimeric LANA:E2C. Neither chimeric p53 nor chimeric c-Myc proteins supported plasmid segregation in partitioning assays (Fig. 8B). These data demonstrate that specific TAD interactions with the transcription machinery are very important for plasmid segregation. We propose that specific interactions with cellular proteins and/or specific subnuclear localization is necessary for plasmid distribution and proper segregation-competent complex formation.

VP16 TAD functionality is gained without Brd4 interaction.

It was recently shown that BPV1 E2 interacts with the double-bromodomain protein Brd4 via its N-terminal TAD. It was proposed that Brd4 is a receptor for E2 on host mitotic chromosomes (74). Subsequent studies demonstrated that Brd4 is also an important component of E2-mediated transcriptional regulation (24, 44, 54, 73). Previous studies demonstrated that the E2 TAD amino acids R37 and I73 are essential for Brd4 binding (9, 54, 56). We showed that mutation of arginine to alanine at position 37 destroys E2 segregation function (2). However, as shown in Fig. 8A, addition of the VP16 TAD to the hinge region of the R37A protein restored segregation function. We were curious as to whether restoration of segregation was due to restored Brd4 binding. We performed an immunoprecipitation assay to assess Brd4 binding (Fig. 8C). As expected, the wt E2 protein was efficiently immunoprecipitated with rabbit anti-Brd4 antibody (Fig. 8C, lane 7). The R37A mutant and chimeric R37A:VP16 were not coimmunoprecipitated (Fig. 8C, lanes 9 and 10), indicating that the VP16 TAD does not restore Brd4 binding. These data indicate that VP16 TAD-restored segregation does not require Brd4 binding, and like the case for HPV, E2 segregation complex formation is Brd4 independent.

DISCUSSION

Papillomaviruses and a number of herpesviruses maintain their genomes as extrachromosomal genetic elements in latently infected proliferating cells. It is thought that attachment of specific viral proteins to host cell chromatin and tethering of the viral genome by the same protein provide the basic mechanism for segregation/partitioning of viral genomes into daughter cells during mitosis. Viral segregation proteins bind to mitotic chromatin with a high affinity, and correct binding is ensured by specific interactions between segregation protein domains and multiple chromatin-bound proteins. Effective tethering of the viral genome to mitotic chromatin occurs via interactions between the segregation protein and multimeric binding sites in the viral genome. Our previous data showed that E2 binding to the BPV1 genome is dependent on at least five E2 binding sites within the BPV1 genome (51, 58). These data indicate that multiple segregation proteins form a complex to ensure stable tethering of the viral genome to mitotic chromatin.

The N-terminal TAD of E2 is required for interactions with replication and transcription activation machineries, and the domain is necessary and sufficient for E2 chromatin attachment (8). The E2 structural/functional domains have been mapped by point mutation analyses (2, 16, 75). From these studies, it is obvious that E2 differs in structural integrity from LANA1 and EBNA1, in that no linear peptide sequence that is capable of binding to mitotic chromatin has been identified. Therefore, it has been difficult to identify the E2 receptor(s) and to study the role of different interactions involved in E2-mitotic chromatin binding. Several linear independent chromosome binding domains have been mapped in EBNA1. The domains may be important for chromosome attachment via interactions with histone H1 (41) or the cellular protein EBNA1 binding protein 2 (EBP2) (57, 72). LANA1 chromosome binding regions have been identified in both the N- and C-terminal protein regions, and these regions are thought to bind cooperatively to mitotic chromosomes to ensure efficient viral genome segregation (6, 29, 30, 34, 52). The first 22 amino acids are sufficient for LANA1 N-terminal chromatin attachment (6, 7, 52). In this study, we used the N-terminal region of LANA1 for complementation studies to investigate E2 chromatin attachment function.

Chimeric LANA:E2C, in which we replaced the E2 TAD with the chromatin attachment domain from LANA1, displayed DNA binding activity in band shift assays, was attached to the mitotic chromatin, and supported plasmid tethering to host chromosomes in metaphase and anaphase (Fig. 2 and 3). However, the chimeric protein was unable to support plasmid partitioning in dividing cells (Fig. 4). Our data indicate that plasmid attachment was not sufficiently strong and failed in late mitosis. Importantly, we observed that the ability of LANA:E2C to cooperatively bind dimeric E2 binding sites did not differ from that of E2C and was considerably lower than that of E2 (Fig. 2C). This means that effective occupation of multimeric E2 binding sites requires very high concentrations of LANA:E2C (compared to those for E2) for effective tethering during mitosis. We studied various concentrations of LANA:E2C and E2 and assessed mitotic chromosome attachment. By staining with E2-specific monoclonal antibodies, we saw that both proteins attached to mitotic chromosomes (Fig. 3). Because the proteins carry identical DBD epitope sequences, we used MAbs 3E8 and 5H4 to measure dot intensities, which are indicative of E2 molecule numbers. Indeed, large-scale high-content image analysis of thousands of mitotic cells revealed that low LANA:E2C concentrations showed lower-intensity staining on mitotic chromosomes, indicating that there were fewer molecules in every dot than the number for E2 (Fig. 6A). At higher LANA:E2C concentrations, LANA:E2C dot intensities were equal to E2 dot intensities (Fig. 6A). These data indicate that LANA:E2C segregation defects observed in our initial experiments may have resulted from incomplete formation of the segregation-competent complex and that formation of fully functional segregation complexes requires high LANA:E2C protein levels. We increased LANA:E2C expression by adding an hEF1α intron sequence to the LANA:E2C expression cassette, which led to a 4-fold increase in LANA:E2C expression (Fig. 7A). The increase in expression was sufficient to rescue segregation/partitioning. We concluded that formation of segregation-competent complexes depends on a certain protein level threshold and that multiple E2 molecules are present in a single segregation complex. In support of this model, iLANA:E2C expression led to characteristic rapid decreases in d1EGFP-positive cell numbers at later time points (Fig. 7B). The GTU plasmids were not capable of replication; thus, after multiple rounds of cell division, there was a reduction of the plasmid copy number in each cell. Most probably, the plasmid copy number was additionally reduced by DNA degradation. The reduction of plasmid copy number was confirmed by a lower d1EGFP fluorescence intensity per cell in the later time points of the segregation assay (data not shown). These events led to a reduction in LANA:E2C expression, which resulted in incomplete assembly of the segregation/partitioning complex.

The cervical keratinocyte cell line W12 accurately models cervical neoplastic progression (48, 61). W12 cells show a spontaneous transition from containing only episomal HPV16 to containing only integrated HPV16. This transition is believed to be a key event in cervical cancer progression (4, 48). In addition to viral genome integration, a progressive reduction in E2 protein expression and a loss of episomal HPV16 plasmids have been observed (49). Pett et al. speculated that HPV-related carcinogenesis must include not only viral integration but also the steps leading to the loss of episome-mediated inhibition of selectable integrants (49). Our data show that formation of the E2 segregation-competent complex functions in a concentration-dependent manner and that a reduction in E2 level effectively abolishes extrachromosomal maintenance of the viral genome. Thus, sufficient E2 expression levels in infected cells are important for stable viral genome maintenance and for regulation of viral oncoprotein expression in integrated genomes.

Mapping of E2 protein functional determinants revealed that the TAD is required for chromatin attachment and E2-mediated plasmid or viral genome segregation. A single amino acid substitution in the TAD can lead to a failure of chromatin association or segregation complex formation, indicating that the structural integrity of the TAD is crucial for these activities (2, 9, 75). It is interesting that the replication activation function of E2 and the interaction with E1 protein are clearly separable from other E2 activities (2, 3, 9, 16, 71). Interestingly, E2 transactivation, chromosome attachment, and plasmid segregation activities are not exclusive. Furthermore, transcriptional activation of mutated E2 proteins tends to correlate with chromosome attachment and plasmid segregation abilities (2, 9). To identify the respective determinants, we constructed chimeric E2 proteins that allowed us to clearly discriminate chromatin attachment and transactivation activities. We added the LANA1 chromosome binding domain or the VP16 TAD from HSV1 to E2 mutants in order to complement different functions. Molecular genetic studies of VP16 have shown that the C-terminal 78 amino acids of VP16 constitute the TAD when linked to a heterologous DBD and that the VP16 TAD is a very strong DNA-binding-site-dependent transcription activator (12, 53). Addition of the VP16 TAD to an E2 mutant that was defective in transcription (R37A mutant) restored transactivation function in this protein. Moreover, addition of the VP16 TAD also restored segregation/partitioning function (Fig. 8A). Alternatively, addition of the LANA1 N-terminal amino acids to E2 mutants that were incapable of binding chromatin but capable of transcriptional activation (R39A and E68A mutants) restored segregation function (Fig. 8A). The R37A mutant, which was capable of binding chromatin, was not functional in segregation/partitioning when the LANA1 domain was added (Fig. 8A). These data clearly demonstrate that the E2 TAD carries two determinants of segregation/partitioning, belonging to two genetic complementation groups. These functions can be complemented by heterologous domains with similar or identical functions. Additional support for this conclusion came from a chimeric molecule in which the E2 TAD was replaced by LANA1 and VP16 domains. The molecules attached to host chromosomes, activated transcription from E2-dependent promoters, tethered plasmid DNA to host chromosomes, and, most importantly, supported plasmid segregation in partitioning assays (Fig. 9). Interestingly, the chimeric LANA:VP16:E2C protein was always less efficient at partitioning than E2 (Fig. 9E). Large-scale image analysis of LANA:VP16:E2C showed that segregation-competent complex formation was not as efficient as that in iLANA:E2C- and E2-transfected cells. However, at higher LANA:VP16:E2C concentrations, E2-like complex formation was detected (data not shown). It is apparent that LANA:VP16:E2C and E2 have different targets on host chromosomes. The N terminus of LANA1 has been shown to interact with H2A-H2B (7). An interaction between E2 and histone proteins has not been identified. Interactions between the E2 protein and the transcription machinery most likely differ from VP16 TAD-transcriptional machinery interactions. The acidic carboxy-terminal region of VP16 can target many general transcription factors and chromatin-modifying coactivator proteins, including TBP (63), TFIIA (31), TFIIB (40), human cofactor PC4 (18, 33), CBP (21, 22), and p300 (32, 35). EBNA1, LANA1, and E2 participate in transcription activation, and they all interact with several cellular proteins that are present in transactivation complexes. All of these viral proteins also interact with Brd4 (39, 47, 54, 74), and Brd4 is involved in E2-mediated transactivation (24, 44, 54, 56). Moreover, Jang and associates showed that E2 and Brd4 were bound to most transcriptionally active promoters in C33A cells (26). Our lab demonstrated by biochemical fractionation that there are two pools of E2 molecules in the cell nucleus, one that localizes to transcriptionally inactive compact chromatin and another that compartmentalizes to transcriptionally active nuclear structures (37). Thus, it is reasonable to assume that Brd4 is the “guiding molecule” during segregation. Our results showed that Brd4 does not interact with the VP16 TAD (Fig. 8C). Similar results were previously described by Senechal et al. (56). We concluded that VP16 TAD restoration of segregation does not require Brd4 interactions, indicating that interactions with other transcription factors or complexes are involved in E2-mediated segregation.

Chimeric molecules containing the TADs from p53 or c-Myc did not support plasmid segregation (Fig. 8B), although the activation domains did restore E2-dependent transcriptional promoter activation (data not shown). These data indicate that interactions with specific complexes during mitosis are required for proper plasmid segregation. This is supported by the fact that a point mutant (F442A) of VP16, in conjunction with the LANA chromatin attachment domain, did not support plasmid segregation (Fig. 9E). The phenylalanine at position 442 plays a crucial role in VP16 transactivation and binding to the TFIID complex (13, 25). Our results showed that the F442A mutant was approximately 4-fold less active in E2-dependent promoter activation than native VP16, which has remarkable transactivation abilities (Fig. 9F). On the other hand, the failure of LANA:F442A:E2C in partitioning (Fig. 9E) shows that this mutation probably inhibits specific interactions with the transcription machinery during mitosis or segregation complex formation that are required for segregation/partitioning.

In conclusion, our results show that the E2 viral segregation protein forms a functional segregation complex via dual interactions with cellular proteins. First, the E2 segregation protein must bind to specific chromatin structural proteins, to which binding is necessary but not sufficient for segregation. Second, formation of the segregation-competent complex requires interactions of multiple E2 molecules and several cellular proteins, including certain components of the transcriptional machinery. Deficiencies in the second activity are complemented by increased protein concentrations, which indicates that the functional segregation complex incorporates several copies of the viral protein.