INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The baculovirus Autographa californica multinucleocapsid nucleopolyhedrovirus (AcMNPV) has been widely used as an eukaryotic expression vector (27) and has been engineered as an alternative biopesticide (41). More recently, baculovirus has been investigated as a potential gene transfer vector for human gene therapy (30). All of these applications rely on the genetic manipulation of the viral genome, usually in the presence of prokaryotic plasmid DNA. AcMNPV contains a closed-circular, double-stranded DNA genome (134 kb) that is tightly associated with a highly basic protein to form a chromatin-like structure (35, 40). Eight A+T-rich, homologous regions (hr) (hr1, hr1a, hr2, hr3, hr4a, hr4b, hr4c, and hr5), interspersed along the genome, were predicted to function as viral origins of replication (7). Utilization of an infection-dependent replication assay led to the identification of hr as putative origins of viral DNA replication (1), supporting this hypothesis. However, other regions on the AcMNPV genome also act as viral infection-dependent replicating sequences (15, 21, 42) so it is still unknown whether each or any of the identified replicating sequences also functions as a DNA replication origin in vivo. For instance, deletion of hr5 from the AcMNPV genome has no apparent effect on virus replication (32). Nevertheless, plasmid replication in virus-infected cells is strictly dependent upon the presence of specific viral sequences such as hr (22, 31), regions within the HindIII K fragment (16, 20), or viral early gene regions (42). Although it is believed that the viral replication machinery specifically interacts with cognate replication origins, there is evidence to suggest that initiation of early viral gene expression and viral DNA replication may be coupled (42).

The demonstration that purified viral genomic DNA is infectious when transfected into insect cells (6) has been exploited for the production of recombinant baculoviruses (34). In addition, cotransfection of insect cells with AcMNPV DNA and a transplacement vector or a reporter plasmid is routinely used in generating recombinant baculoviruses or in measuring the level of reporter gene expression or reporter replication. However, the consequences of plasmid replication on these applications have not been investigated. It has been observed that pUC-based plasmids without baculovirus inserts replicate when cotransfected with viral DNA into insect cells (11, 17). The basis of this plasmid DNA replication is unknown, although it has been speculated that it may result from the acquisition of hr sequences following cotransfection (14). We thought that these results might reflect differences in the specificity of plasmid replication in cotransfected versus infected cells so we have investigated this phenomenon further. An understanding of the nature of plasmid replication in cotransfected cells may shed light on the specificity and mechanism of baculovirus DNA replication. In this report, we describe bacterial plasmid DNA replication in insect cells and its relationship with the viral DNA replication process. The data demonstrate that plasmid replication is independent of the presence of viral DNA sequences in cis but is initiated by the viral replication machinery. The replicated plasmid DNA has a concatenated structure that can be incorporated into the viral genome where it is packaged into mature virions and continuously passaged with the viral DNA following subsequent rounds of replication.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and virus. Spodoptera frugiperda (Sf21) cells were maintained by passage in TC-100 medium supplemented with 10% fetal calf serum. AcMNPV (strain HR3) was prepared and titrated as previously described (9). Extracellular budded virus was purified from infected cell supernatants through 25 to 65% sucrose gradients as previously described (36). DNA was prepared from purified virions by standard protocols (38).

Plasmids. All plasmids were propagated in Escherichia coli DH5 cells and purified on Qiagen Tip 500 columns following the manufacturer's instructions (Qiagen Inc.). The AcMNPV HindIII Q fragment was cloned into the HindIII site of pUC18 to generate pAchr5.

DNA cotransfection and replication assays. Sf21 cells were cotransfected by mixing plasmid DNA with AcMNPV DNA and 10 µl of Lipofectin (Gibco-BRL) in a total volume of 75 µl and then adding the mixture to washed cell monolayers (10⁶ cells per cotransfection assay). After 6 h of incubation at 28°C, the DNA-Lipofectin mixture was removed, the cells were washed twice, overlaid with fresh medium, and incubated at 28°C for 48 h. Total intracellular DNA was purified, and the ability of plasmid DNA to replicate in Sf21 cells was monitored by its differential susceptibility to DpnI digestion as previously described (23). The relative amounts of viral and plasmid DNA present in total intracellular DNA preparations were determined by densitometry of the hybridization signals of calibrated amounts of plasmid and viral DNA in 1 µg of either total intracellular DNA or purified viral DNA by using the public domain computer program NIH Image (version 1.54). In some cases, DNA concentrations were determined by phosphorimaging analysis of Southern blots.

Subcloning of integrated plasmid DNA. About 2 to 3 µg of passage-three (P3)-purified budded virus DNA was digested to completion with MluI, and then the restriction enzyme was inactivated by incubation at 85°C for 20 min. ATP was added to a final concentration of 1 mM, and the mixture was incubated in the presence of T4 DNA ligase for 12 h at 16°C. The ligation products were desalted and then transformed into competent E. coli DH5 with an ElectroPorator (Invitrogen) under the conditions recommended by the manufacturer. The transformants were selected in the presence of 100 µg of ampicillin/ml.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmid replication is independent of specific viral sequences. To confirm that superhelical plasmid DNA lacking any baculovirus sequence could replicate in the presence of baculovirus DNA, plasmid pUC18, pBSK, or pBR322 DNA was cotransfected with viral DNA into Sf21 cells, and total intracellular DNA was analyzed at 48 h postcotransfection by digestion with DpnI. All three plasmids replicated in the presence of viral DNA almost as efficiently as the reporter plasmid carrying an hr (pAchr5) (Fig. 1, lanes 5 to 12). Plasmid replication was dependent upon the presence of viral DNA since pUC19 did not replicate in the absence of viral DNA in the transfection mixture (Fig. 1, lane 3). Because it has been clearly shown that plasmid DNA replication is absolutely dependent upon the presence of certain viral DNA sequences in cis when plasmid transfected cells are postinfected with virions (14), these results suggested that initiation of plasmid DNA replication in cells cotransfected with viral DNA was different than that in infected cells.

View larger version (44K): [in this window] [in a new window]   FIG. 1.   Relative replication efficiency of plasmid DNA in cotransfected insect cells. Sf21 cells (106) were transfected with 0.5 µg of AcMNPV DNA (lanes 1 and 2) or 0.5 µg of pUC19 DNA (lanes 3 and 4) or were cotransfected with 0.5 µg of AcMNPV DNA plus equal molar equivalents of pAchr5 (lanes 5 and 6), pUC19 (lanes 7 and 8), pBSK (lanes 9 and 10), or pBR322 (lanes 11 and 12) DNA. Total intracellular DNA was purified at 48 h posttransfection and digested with EcoRI (DpnI) or EcoRI plus DpnI (DpnI+). After electrophoresis, the DNA fragments were transferred to a Qiagen nylon membrane and hybridized with 32P-labeled pUC18 DNA. The positions of marker DNA (sizes in kilobases) are indicated on the left of the blot.

Structure of the replicated plasmid DNA in cotransfected cells. It has been suggested that baculovirus may use a rolling-circle mechanism to replicate its genome because plasmids containing hr replicated into high-molecular-weight concatemers in virus-infected cells (22) and multimers of viral DNA have been detected in virus-infected cells (29). In order to determine whether these criteria extended to plasmids that did not carry any viral sequences, we studied the structure of the replicated plasmid DNA by restriction digestion of total intracellular DNA from cells cotransfected with pUC19 and viral DNA. HindIII or PstI digested the replicated plasmid DNA (DpnI resistant) into a 2.7-kb fragment that comigrated with the EcoRI-digested input pUC19 (Fig. 2A, lanes 3 to 5 [left]). When digested with only DpnI, the replicated plasmid DNA migrated to the position of high-molecular-weight DNA, similar in size to undigested viral genomic DNA (Fig. 2A, lane 6 [left]). Because these gels were not capable of resolving large DNA molecules, the comigration of high-molecular-weight plasmid DNA and viral DNA was confirmed by pulsed-field gel electrophoresis (data not shown). In comparison, undigested input plasmid DNA migrated as supercoiled and relaxed forms when undigested (Fig. 2A, lane 2 [left]) or as small fragments when completely digested with DpnI (data not shown). These results suggest that the plasmid DNA was replicated into high-molecular-weight concatemers. To confirm this hypothesis, total intracellular DNA was completely digested with DpnI and partially digested with SmaI (Fig. 2B). Distinct fragments of 2.7, 5.4, 8.1, and 10.7 kb as well as a high-molecular-weight smear were detected, indicating that multimers of replicated plasmid DNA were formed during replication (Fig. 2B). Quantification of the hybridization signals obtained from calibrated amounts of purified plasmid and viral DNA in 1 µg of cellular DNA revealed that for every microgram of viral DNA, there was 400 to 500 pg of replicated plasmid DNA in the cotransfected cells.

View larger version (53K): [in this window] [in a new window]   FIG. 2.   Replicated plasmid in cotransfected Sf21 cells exists as high-molecular-weight DNA. (A) Sf21 cells were transfected with 0.5 µg of AcMNPV DNA (lane 1), 0.5 µg of pUC19 DNA (lanes 2 and 3), or cotransfected with 0.5 µg of AcMNPV and pUC19 DNA (lanes 4 to 9). Total intracellular DNA was purified at 48 h postcotransfection and digested with DpnI plus HindIII (D+H, lane 4), DpnI plus PstI (D+P, lane 5), DpnI alone (D, lane 6), DpnI plus NotI (D+N, lane 7), DpnI plus EagI (D+Ea, lane 8), or DpnI plus MluI (D+M, lane 9). One microgram of purified, undigested input AcMNPV DNA (U, lane 1), 200 pg of undigested pUC19 DNA (U, lane 2), or 200 pg of pUC19 DNA digested with EcoRI (E, lane 3) were included as controls. After electrophoresis, DNA samples were blotted and probed with 32P-labeled pUC18 DNA (left). After the exposure shown on the left, the membrane was stripped and reprobed with 32P-labeled AcMNPV DNA to confirm the positions of viral DNA fragments (right). The positions of marker DNA (sizes in kilobases) are indicated between the blots. (B) Total intracellular DNA, purified at 48 h posttransfection from cells cotransfected with 0.5 µg of AcMNPV DNA and 0.5 µg of pUC19 DNA, was completely digested with DpnI (lane 1) and then partially digested with increasing amounts of SmaI (lanes 2 to 9). The fragments were separated by electrophoresis and then blotted and probed as described in the legend for Fig. 1. The positions of size markers (kilobases) are indicated on the left of the blot while the positions of various forms of concatenated plasmid DNA are indicated on the right.

Integration of replicated plasmid DNA into viral genomes. The presence of DpnI-resistant plasmid DNA, in the absence of digestion with any other enzymes, comigrating with undigested viral genomic DNA (Fig. 2A, lane 6) suggested that some pUC19 DNA molecules might be integrated into the viral genome. To test this possibility, total intracellular DNA from cotransfected cells was digested with DpnI and NotI, EagI, or MluI (enzymes that digest viral genomic DNA but not pUC19 DNA). If replicated pUC19 DNA was integrated, digestion of viral DNA would release fragments carrying plasmid DNA and the released DNA would have a different size than its integrated form. On the other hand, if the replicated plasmid DNA was not covalently integrated, digestion of viral DNA would not affect the mobility of pUC19 DNA. The results revealed that the replicated plasmid DNA migrated as a smear ranging from about 6 to more than 20 kb after digestion with NotI or from 4 to 19 kb when digested with EagI or MluI (Fig. 2A, lanes 7 to 9). These results suggest that significant amounts of replicated plasmid DNA were integrated into the viral genomic DNA. Since no particular band of plasmid DNA was detected after digestion, either the integration occurred at multiple sites around the viral genome or the integrated pUC19 DNA existed in different sizes or both.

View larger version (46K): [in this window] [in a new window]   FIG. 3.   Detection of replicated plasmid DNA in progeny budded virus particles. Sf21 cells (106) were cotransfected with 0.5 µg of AcMNPV DNA plus 0.5 µg of pUC19 DNA (lanes 3 to 11) or pAchr5 (lane 12). Progeny virions were harvested at 72 h postcotransfection (lanes 3 and 12). The supernatant from the pUC19-transfected cells was serially passaged undiluted four times (lanes 4 to 7) or passaged with different amounts of viruses (lanes 8 to 11). Budded virions from each passage supernatant were purified by sucrose gradient centrifugation, and the virion DNA was purified and doubly digested with SmaI plus DpnI (lanes 3 to 12). Following agarose gel electrophoresis, the fragments were blotted onto a Qiagen nylon membrane and hybridized with 32P-labeled pUC18 probe. The positions (in kilobases) of SmaI-digested pUC18 (lanes 1) and pAchr5 (lane 2) DNA, included as the molecular weight markers and hybridization controls, are indicated on the left. The arrows on the right indicate the positions of SmaI-linearized plasmid DNA (pUC19 or pAchr5) contained in the viral DNA and the fragments of the digested pUC19 or pAchr5 DNA likely covalently linked to viral DNA (high-molecular-weight DNA).

View larger version (62K): [in this window] [in a new window]   FIG. 4.   Conformation of plasmid DNA packaged into budded virions. P3 budded virions from serially passaged virus stocks originally obtained from cells cotransfected with pUC19 and AcMNPV DNA were purified, and the virus DNA was isolated. Samples of undigested virion DNA (U, lane 6) or virion DNA digested with EcoRI (E, lane 4) HindIII (H, lane 5), PstI (P, lane 6) NotI (N, lane 8), EagI (Ea, lane 9), MluI (M, lane 10), Sse8387I (Se, lane 11), or Sse8387I plus MluI (Se+M, lane 12) were resolved on agarose gels and Southern blots were prepared. Controls included undigested purified AcMNPV DNA (AcMNPV, lane 1) and undigested pUC19 DNA (U, lane 2) or pUC19 DNA digested with EcoRI (E, lane 3). The blots were first probed with pUC19 DNA (left) and then stripped and reprobed with AcMNPV DNA (right).

Sequence of integration junction sites. These data clearly demonstrated that plasmid sequences had integrated into the viral genome because of coreplication. To investigate the location and specificity of the integration site(s) on the AcMNPV genome, we developed a strategy to clone regions of viral DNA which contained junctions between viral and plasmid DNA. P3 virion DNA (Fig. 3, lane 5) was digested with MluI, which restricted viral but not pUC19 DNA (Fig. 4, lane 10), to release DNA fragments including those potentially carrying plasmid DNA. The digestion fragments were treated with T4 DNA ligase and selected for the presence of plasmid DNA by transforming E. coli DH5 cells. Any fragments retaining the ampicillin resistance gene and the plasmid origin of replication were selected in the presence of ampicillin. Fifty recombinant colonies were selected, and the plasmid DNAs were purified and mapped by digestion with HpaII to localize the viral insert within the remaining plasmid sequences (Fig. 5). By comparison with pUC19 cut with HpaII, 22 of the 50 clones revealed a single missing (or possibly shifted) HpaII fragment (4 were missing A, 4 were missing B, 4 were missing C, 8 were missing D, and 2 were missing H). Another three clones were missing two bands, either the HpaII A and D (pM5), B and H (pM7), or D and H (pM25) bands. Therefore, in these clones, the junction sites between pUC19 DNA and AcMNPV DNA were likely located within these specific regions of pUC19. The junction sites for the other 25 clones could not be determined by this approach because no obvious changes in the HpaII fragment patterns were detected. It is likely that the integrated plasmid DNA in these clones was multimeric, resulting in multiple copies of all HpaII fragments obscuring the presence of viral DNA in a specific HpaII fragment region.

View larger version (84K): [in this window] [in a new window]   FIG. 5.   Restriction mapping of recombinant plasmids with HpaII digestion. Purified P3 viral DNA carrying integrated plasmid DNA was digested with MluI, treated with T4 DNA ligase, and transformed into DH5 cells to selectively amplify viral sequences carrying pUC19 DNA. Recombinant clones were selected in the presence of ampicillin, and purified plasmid DNA from these clones was digested with HpaII and separated on 5% nondenaturing polyacrylamide gels. The migration positions and the relative sizes of pUC19 HpaII fragments are indicated on the left and right, respectively. The missing HpaII fragment(s) detected in each lane, indicating the location of an insertion of viral DNA, is shown at the bottom of each figure for certain clones.

View larger version (52K): [in this window] [in a new window]   FIG. 6.   DNA sequence analysis of the junction sites between pUC19 and AcMNPV DNA. (A) The HpaII map of pUC19 is shown at the top, and below are maps of recombinant plasmids that were predicted to carry viral inserts. Downward arrowheads indicate the approximate locations of viral inserts on these plasmids. The dashed lines indicate regions that were sequenced to determine the junction sites between pUC19 and viral DNA. Arrows indicate the specific primers and directions of the sequencing reactions. (B) Each nucleotide sequence is presented as a continuous sequence in the direction of 5' to 3' from either end of a specific insert. The name of each plasmid and the corresponding primers used for sequencing are indicated at the beginning of the sequence. pUC19 sequences are in italics, and viral DNA sequences are underlined. The numbers above the sequences indicate the nucleotide positions on pUC19 (43) or the AcMNPV genome (2). The bold letters represent insertion elements that do not align with either pUC19 or viral DNA sequences. The numbers between slashes represent numbers of continuous nucleotides which are not shown due to space limits.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this report, we confirm previous reports that plasmids, when cotransfected with baculovirus DNA into insect cells, can replicate even when no putative viral origin of DNA replication in cis is present on the plasmid (11, 13, 17). These data are in full agreement with our previous results which demonstrated that nonspecific plasmid replication took place even in the presence of virus infection, if the infection occurred within about 4 h of plasmid transfection (42). In addition, our data are consistent with previous observations that the baculovirus DNA replication machinery supported the replication of plasmids carrying the simian virus 40 (SV40) origin in infected insect cells expressing the large T antigen (25). Apparently, when SV40 T antigen was synthesized, it recognized the SV40 origin on plasmid DNA, and replication, coupled with recombination, occurred in the presence of the baculovirus replication machinery. Together, these data suggest that during the early period of baculovirus infection, input DNA may be compartmentalized in a manner that could allow nonspecific initiation to occur on any naked DNA once the viral proteins for DNA replication are synthesized and transported to these nuclear sites.

The replicated plasmid DNA detected in our experiments was concatenated, could be integrated into viral genomes at multiple locations with no particular sequence specificity, and was packaged with viral genomic DNA into budded virions. This extends previous studies where hr-directed virus-dependent plasmid DNA replication revealed concatemer formation of plasmid DNA following replication but without any evidence of integration or packaging of plasmid DNA (22). Our data are consistent with a rolling-circle mechanism for replication of plasmid molecules resulting in the formation of multimeric molecules that then could be integrated into the viral genome. These multimeric inserts could also be formed either by plasmid-plasmid recombination followed by integration into a viral genome or by unequal crossing-over between two AcMNPV genomes that already contained plasmid inserts. Homology-independent packaging of head-to-tail multimers of pBR322 DNA into T4 phage particles has been detected by a transduction assay (37). Transduction of pBR322 did not occur when recombination-negative mutants were used, suggesting that recombination was required for this process to occur. The homology-independent packaging detected in our experiments was probably the result of illegitimate recombination between plasmid DNA and one or more sites on the AcMNPV genome. Because the only viral gene products required for replication and integration of plasmid DNA were those also essential for viral DNA replication (data not shown), the same minimal set of virus-encoded enzymes function during both recombination and replication, although cellular enzymes such as topoisomerases, ligase, and nucleases are probably involved in these two processes.

By comparing the percentages of the replicated plasmid DNA contained in progeny virions with the amounts found in the cotransfected cells, the incorporation rate of replicated plasmid DNA into virions was estimated to be 10 to 25% of the total replicated plasmid DNA. Although it is possible that some of the plasmid DNA was packaged into virion-like particles without integration, the presence of integrated plasmid DNA suggests a high degree of involvement of illegitimate recombination during DNA replication supported by the baculovirus replication machinery. These data are consistent with observations that the bacterial transposon Tn5, when inserted into the baculovirus genome as an indicator for recombination, exhibits high levels of Tn5 inversion (26). Our results suggest that while baculovirus may use a rolling-circle mode to replicate its DNA, resulting in concatemeric DNA as shown for the replicated plasmid DNA, recombination is also involved. This is reminiscent of herpes simplex virus type 1 DNA replication where, although concatemeric forms of replicated viral DNA can be easily detected, newly replicated viral DNA is composed of highly branched, complex networks (3, 33, 44). Thus, there is increasing evidence to show that the successful growth of large DNA viruses such as T4 and herpes simplex virus requires both DNA replication and recombination (28, 39). Our data suggest that recombination also plays an important role during baculovirus replication.

Analysis of the integration sites revealed that plasmid DNA was linked with different regions of the viral DNA, some of which were separated by as much as 50 kb. These data suggest that nonhomologous recombination between plasmid DNA and viral DNA could lead to deletion of large portions of the viral genome and possible amplification of multimers of plasmid DNA. If there are certain limits to the amount of DNA that can be packaged within the virion, the packaging of concatenated plasmid DNA might be limited by its ultimate length and the presence of specific packaging signals. The genomes of some defective baculoviruses contain large deletions of the viral genome (5, 18), although the arrangement of this DNA within virions has not been investigated. A similar mechanism of nonhomologous recombination could be responsible for the generation of these defective genomes in both cases. Our data also suggest that integration is unlikely to rely on any specific target sequence since no particular pattern of sequence was observed at any of the junction sites analyzed, either on the plasmid or the viral DNA. These results support the concept that the integration of plasmid DNA into the viral genome is a result of random nonhomologous recombination.

Plasmid replication, recombination, and integration in cotransfected cells may suggest a way by which baculoviruses have acquired nonhomologous DNA sequences from host cells during evolution and may also account for the differences in gene content among various baculovirus species (19). For instance, the viral genes encoding proliferating cell nuclear antigen (PCNA) and ubiquitin may derive from the cellular homologues by such a mechanism. It has been clearly shown that cellular repeated sequences have inserted into the AcMNPV 25K gene region at a high frequency, generating few polyhedra (FP) mutants (4, 10). In this case, the 25K gene mutations were easily detectable because of obvious changes in polyhedron morphogenesis. If more selection markers were available, more integration events would likely be revealed. For example, in the polyhedra morphology mutant M5, two identical host cellular repetitive sequences of 290 bp were found inserted at the 2.6 and 46 map unit regions (5). These results suggest that cellular sequences as well as plasmid sequences can insert at multiple locations in the viral genome.

The insertion elements contained in the clones pM9, pM22, and pM25 as well as pM7 aligned with sequences from different organisms including humans, Caenorhabditis elegans, Drosophila melanogaster, and yeast. Some of these aligned sequences were repeated sequences, such as telomeric-repeat-like internal eliminated sequence (12), or sequences from mitochondrial DNA (8), suggesting but not proving that these unknown insertion sequences were likely derived from the host insect cells. Insertion of unknown sequences has also been observed in the genome of defective viruses (5, 18). The integration of certain cellular sequences may supply an evolutionary advantage to baculoviruses replication. Our results suggest that the baculovirus DNA replication mechanism can be promiscuous in template choice and/or recombination partners. The fact that cotransfected DNA was replicated and integrated into the viral genome could also explain the occurrence of unstable recombinant baculoviruses that cannot be plaque purified, for example while trying to identify essential viral genes by gene insertion (24). Our results are also relevant to the safety aspect of using genetically engineered baculoviruses as biopesticides or delivery vectors for gene therapy since many of these agents are constructed by cotransfection of viral and plasmid DNAs. In addition, if cellular homologues of oncogenes or retrovirus-like repeated sequences integrated into the genome of baculoviruses during scale-up, these elements might be persistently maintained in the population of viruses as demonstrated in this study. The implications and consequences of the integration of these foreign elements would certainly deserve close attention.