Genetic Requirements for Homologous Recombination in Autographa californica Nucleopolyhedrovirus

Abstract

It is known that baculovirus infection promotes high-frequency recombination between its genomes and plasmid DNA during the construction of recombinant viruses for foreign gene expression. However, little is known about the viral genes necessary to promote homologous recombination (HR). We developed an assay to identify viral genes that are necessary to stimulate HR. In this assay, we used two plasmids containing extensive sequence homology that yielded a visible and quantifiable phenotype if HR occurred. The plasmids contained the green fluorescent protein gene (gfp) that was mutated at either the N or the C terminus and a viral origin of DNA replication. When the plasmids containing these mutant gfp genes were transfected into insect cells alone or together, few green fluorescent protein (GFP)-positive cells were observed, confirming that the host cell machinery alone was not able to promote high levels of HR. However, if viral DNA or viral genes involved in DNA replication were cotransfected into cells along with the mutant gfp-containing plasmids, a dramatic increase in GFP-positive cells was observed. The viral genes ie-1, ie-2, lef-7, and p35 were found to be important for efficient HR in the presence of all other DNA replication genes. However, ie-1 and ie-2 were sufficient to promote HR in the absence of other viral genes. Recombination substrates lacking a viral origin of replication had similar genetic requirements for recombination but were less dependent on ie-1. Interestingly, even though HR was stimulated by the presence of a viral origin of DNA replication, virally stimulated HR could proceed in the presence of the DNA synthesis inhibitor aphidicolin.

Studies on homologous recombination (HR) among baculovirus genomes in cell culture and in the wild are limited, curtailing our understanding of genetic diversity among baculovirus strains and virus evolution and our ability to design ecologically safe and efficient biological control agents. Recombination of virus genomes within insects results in an increase in virus genetic heterogeneity in wild populations (11- 13, 41, 45, 60) and in cell culture (17, 23, 61, 71). Integration of plasmid DNA by nonhomologous recombination into the baculovirus Autographa californica nucleopolyhedrovirus (AcMNPV) DNA has also been observed in cell culture (69). In addition to shedding light on these processes, identification of the viral genes involved in HR may be useful for studying the function of genes in the host organism by specifically targeting genes in insects infected with baculoviruses or other viruses carrying baculovirus recombination-specific genes. For example, gene targeting mediated by baculoviruses in the silkworm Bombyx mori has been successful (72). Also, knowledge of the mechanism of recombination in the host organism may provide a tool to better develop recombination between baculovirus and plasmid DNA after lipofection of insect larvae. The latter has been suggested and investigated to generate recombinant viruses in the event an insect cell line is not available (18). In cell culture, recombination between plasmids carrying a foreign gene and baculovirus DNA has been exploited to construct recombinant viruses overexpressing a foreign gene; however, the involvement of viral genes in this process had not been investigated until recently. The only study that has examined the requirement of specific baculovirus genes in HR concluded that the AcMNPV DNA replication genes ie-1, lef-1, lef-2, lef-3, p35, and dnapol were necessary for high-frequency HR in the presence of an origin of viral replication, whereas ie-2 and lef-7 had little effect as monitored by the inversion of sequences within a bacterial transposable element (38).

Current evidence suggests that the genome of AcMNPV requires both cis- and trans-acting elements for replication. cis-Acting elements involved in viral DNA replication have been identified by using two main strategies: (i) the analysis of defective virus genomes arising after serial undiluted passage of virus stocks in cell culture (26, 27, 29) and (ii) the characterization of replicated plasmid DNA containing cis-acting elements introduced into infected cells (1, 30, 31, 49-51, 70). Together, these studies have identified homologous regions (hrs) that are distributed throughout baculovirus genomes and other nonhomologous regions as origins of viral DNA replication. Each hr contains several copies of an imperfect palindrome flanked by direct repeats. AcMNPV has eight hrs, each containing between one and eight imperfect palindromic repeats bisected by an EcoRI recognition site.

AcMNPV genes involved in viral DNA replication have been identified by testing the ability of AcMNPV genes involved in late and very late gene expression or identifying genes within overlapping genomic clones to amplify plasmids containing hrs (24, 36). These reports determined that five genes—ie-1, lef-1, lef-2, lef-3, and p143—were required for viral DNA replication but, depending on the time they were analyzed or on the specific hr used, five additional genes—ie-2, lef-7, dnapol, p35, and pe-38—were either required for or stimulated DNA replication.

Most of the AcMNPV genes involved in DNA replication have either sequence homology to other proteins involved in DNA biosynthesis or have been implicated in the process by either characterizing viruses containing temperature-sensitive mutations in the gene or by in vitro activities associated with the process of DNA replication as described below. The genes ie-1, ie-2, and pe-38 are transactivators of other viral genes (5, 33, 48, 67). A virus carrying a temperature-sensitive ie-1 allele is defective in viral DNA replication at the nonpermissive temperature (56). Also, dimers of the IE-1 protein can bind origins of DNA replication (7, 16, 57, 58). The product of lef-1, LEF-1, is a DNA primase, and LEF-2 may have a role as a DNA primase accessory protein (43). LEF-3 has been shown to bind single-stranded DNA in vitro (19), and LEF-7 contains single-stranded-DNA-binding motifs (36), but its activity as a single-stranded binding protein has not been determined either in vitro or in vivo. P143 is a helicase (34, 40), and a virus with a temperature-sensitive mutation in p143 is defective in DNA synthesis at the nonpermissive temperature (4, 14, 34). The DNA polymerase gene of AcMNPV, dnapol, has motifs conserved in other DNA polymerases (64) and has been shown to possess polymerase activity in vitro (20, 39). The DNA polymerase gene was also found to be essential for transient plasmid DNA replication (1, 24, 25, 51), although it only stimulated DNA replication in one report (36). It has been suggested that the host DNA polymerase can work with viral proteins to carry out the biosynthesis of viral DNA (36). The function of p35 in these transient replication assays may be associated with its ability to block apoptosis, possibly induced by ie-1 (52), since other genes that block apoptosis can substitute for p35 (36). Also, the requirement for p35 in late gene expression and DNA replication increases with time as more cells are undergoing cell death (62; E. A. Crouch and A. L. Passarelli, unpublished data).

We developed a simple and visual method to identify specific viral genes that stimulated HR. In this recombination assay, we used two target plasmids containing an overlapping region of sequence homology that yielded a quantifiable phenotype if HR occurred. We found that the AcMNPV genes ie-1, ie-2, lef-7, and p35 were necessary for efficient HR in the presence of the other AcMNPV DNA replication genes and when the gfp homologous sequences were cis-linked to an hr. However, ie-1 and ie-2 alone were sufficient to promote high-frequency HR. In contrast, HR was less dependent on ie-1 in the absence of an hr. Furthermore, AcMNPV genes stimulated HR in the absence of optimal DNA replication.

MATERIALS AND METHODS

Cells and virus.

The cell line IPLB-SF-21 (SF-21) (65) derived from the fall armyworm, Spodoptera frugiperda, was grown at 27°C in TC-100 medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) and 0.26% tryptose broth as previously described (47). AcMNPV L1 strain (28) was propagated in SF-21 cells and used to infect cells in the absence or presence of aphidicolin as previously described (47). Viral DNA was isolated and used for transfections as described elsewhere (47).

Reporter plasmids and plasmid constructions.

Ten reporter plasmids were constructed encoding the enhanced green fluorescent protein (GFP), a red-shifted variant of the wild-type GFP (Clontech Laboratories), or a defective version of GFP under control of the Drosophila heat shock protein 70 (hsp70) promoter. In addition, some of the plasmids contained a fragment of AcMNPV hr5 immediately upstream of the promoter sequences. The control plasmids, pHSGFPhr5 and pHSGFP, contained full-length gfp with or without hr5, respectively. The plasmids pHSGFPNhr5 and pHSGFPN with or without hr5, respectively, contained DNA predicting an N-terminal fragment of GFP corresponding to amino acids 1 to 150. Two plasmids, pHSGFPChr5 and pHSGFPC, with or without hr5, respectively, contained DNA with the C terminus of GFP corresponding to amino acids 77 to the translational stop (codon 240). The latter plasmids lacked a translational start codon. The hr5 fragment in plasmids pHSGFPNhr5 and pHSGFPChr5 was obtained from a 479-bp MluI fragment present in a pUC8 clone containing the HindIII-Q fragment of AcMNPV. Plasmids pHSGFPV2TGAhr5, pHSGFPV2TGA, pHSGFPY152TAAhr5, and pHSGFPY152TAA contained the entire gfp open reading frame but a stop codon was inserted at two different positions within the open reading frame. In pHSGFPV2TGAhr5 and pHSGFPV2TGA, the first guanosine of the second codon, GTG (valine), in the sequence GTGA of GFP was deleted by using the QuikChange mutagenesis kit (Stratagene) creating an opal stop codon (TGA) following the initial methionine. In pHSGFPY15TAAhr5 and pHSGFPY15TAA, the last thymidine in the tyrosine codon (TAT) at amino acid position 152 was changed to an A by using the QuikChange mutagenesis kit, thereby creating an ochre stop codon (TAA). All mutations were confirmed by nucleotide sequencing. The hr5 sequences in pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 were obtained from phcwt (54). All reporter plasmids had a polyadenylation signal derived from the Orgyia pseudosugata inhibitor of apoptosis gene (Op-iap).

A set of nine plasmids containing AcMNPV genes involved in DNA replication—pHSEpiHis-lef-1, pHSEpiHis-lef-2, pHSEpiHis-lef-3, pHSEpiHis-lef-7, pHSEpiHis-ie-1, pHSEpiHis-ie-2, pHSEpiHis-dnapol, pHSEpiHis-p143, and pHSEPIp35—hereafter referred to as replication late expression factors (lefs), have been previously described (55). The plasmid pHSEpiOpiapVI⁺ containing Op-iap and used to substitute for p35 has also been described (66).

DNA cotransfections and enumeration of GFP-positive cells.

Control plasmid (2 μg) or two recombination substrates (1 μg each) and AcMNPV DNA (1 μg) or DNA replication lefs (0.5 μg each) as indicated were introduced into 0.5 × 10⁶ cells plated on 35-mm cell culture dishes by using 3 μl of a 1.5:1 mixture of N-[2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium chloride salt (Avanti Polar Lipids, Inc.) and l-alpha phosphatidylethanolamine (Sigma) lipids. The vector pBlueScript II SK(+) (Stratagene) was used in each cotransfection reaction to maintain a constant concentration of DNA per reaction in each experiment. The control plasmid containing full-length gfp was used to monitor the efficiency of cotransfection in each experiment. Cells transfected with the control plasmid were also scraped and counted in a hemacytometer in a known volume of media in order to determine the total number of cells in each dish and calculate the number of cells that underwent HR. Cells were maintained for 4 h at 27°C in the liposome and DNA mix. This mix was then replaced with TC-100 containing 10% fetal bovine serum and incubated at 27°C.

Approximately 48 h after liposome and DNA were added to cells, GFP-positive cells were counted in a total of 20 fields at random throughout the dish under ×200 magnification (0.916-mm²/field of view). In a typical experiment wherein all of the replication lefs and recombination substrates containing hr5 were cotransfected, we counted between 300 and 2,000 GFP-positive cells in 20 fields of view at ×200 magnification. All experiments were performed independently three to five times.

DNA replication assays.

Cotransfected cells were harvested 48 h after transfection, and DNA was isolated and processed as described previously (8, 36). One-fifth of the extracted DNA was digested with BamHI to linearize the plasmids and DpnI between 8 and 24 h and then resolved by 0.7% agarose gel electrophoresis. DNA was then transferred to a Zeta-Probe membrane (Bio-Rad) and hybridized to a fragment containing the gfp open reading frame previously radiolabeled by a nick translation system (Promega). Blots were visualized by autoradiography. Experiments were done three times using independent transfections.

Aphidicolin inhibition assays.

Approximately 0.5 × 10⁶ cells per 60-mm culture dish were transfected with 1 μg of each pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 or 2 μg of pHSGFPhr5 as specified above. After 24 h, cells were infected with AcMNPV at a multiplicity of infection of 10 PFU per cell. Aphidicolin (Sigma) was added at a final concentration of 5 μg/ml after the 1-h adsorption period, and cells were maintained in medium containing aphidicolin thereafter. For these experiments, GFP-positive cells were counted from 10 fields of view magnified at ×200 at 24 h postinfection, and DNA was extracted for replication assays 24 h postinfection. Results for a total of three independent experiments are reported.

RESULTS

Experimental strategy.

We were interested in defining the specific roles of AcMNPV genes involved in viral DNA replication in HR since it had been reported previously that HR in insect cells was dependent on DNA replication by supplying AcMNPV DNA replication genes in trans and a viral origin of replication in cis (38). To this end, we constructed two control plasmids and four sets of plasmids, each consisting of two HR targets (Fig. 1). The control plasmids contained the full-length gfp under the control of the Drosophila hsp70 promoter either containing a portion of hr5 (pHSGFPhr5) or lacking hr5 (pHSGFP). The first two sets of HR targets were constructed with or without hr5, as noted in the plasmid name, and either C-terminal (pHSGFPNhr5 and pHSGFPN) or N-terminal (pHSGFPChr5 and pHSGFPC) deletions of GFP. N and C terminally truncated GFP constructs contained a region of 222-bp that overlapped to facilitate HR. In order to obtain GFP activity, HR would have to occur in this 222-bp overlapping region. The second two sets of HR targets were also constructed with or without hr5 but contained full-length gfp sequences with an N-terminal (pHSGFPV2TGAhr5 and pHSGFPV2TGA) or a C-terminal (pHSGFPY152TAAhr5 and pHSGFPY152TAA) premature stop codon in the coding sequence. These four constructs shared a larger recombination target sequence (450 bp) that could result in GFP activity than the deleted gfp constructs. None of the gfp defective reporter plasmids produced any detectable GFP activity as demonstrated by the lack of fluorescent cells after transfection of each plasmid DNA alone into SF-21 cells (Fig. 2 and data not shown). These plasmids were then used to determine whether AcMNPV genomic DNA or viral genes could promote HR between the target segments. GFP-positive cells served as an indicator that HR had occurred between the target plasmids. This method allowed both the identification of the specific genes necessary for the process and the quantification of this activity.

FIG. 1.

Plasmid constructs used to assay HR. (A) Full-length gfp constructs under control of hsp70 promoter, pHSGFP, are illustrated on the top line. Constructs containing a C-terminal (pHSGFPN) or an N-terminal (pHSGFPC) deletion and an overlapping region of homology are also shown. Constructs pHSGFPN and pHSGFPC contain amino acids 1 to 150 and 77 to 240 of GFP, respectively. (B) The gfp-containing constructs in which the second amino acid, valine, was changed to an opal stop codon (TGA) (pHSGFPV2TGA) or tyrosine 152 was altered to an ochre stop codon (TAA) (pHSGFPY152TAA) are illustrated, showing the approximate site of the mutation. The “±hr5” indicates that each construct was engineered with (+) or without (−) an hr5 immediately upstream of the promoter region. hsp70p, hsp70 promoter.

FIG. 2.

Stimulation of HR by AcMNPV. (A) GFP-positive cells were counted 48 h after transfection with AcMNPV genomic DNA or the plasmids indicated below each column. The number of GFP-positive cells counted is noted at the top of the bars. This is a single experiment representative of at least three experiments. (B and C) Cells were transfected with the plasmids indicated below each panel and photographed at 48 h posttransfection. In the top panels cells were photographed under visible light, and in the lower panels cells were photographed under fluorescence. V2TGA hr5, pHSGFPV2TGAhr5; Y152TAA hr5, pHSGFPY152TAAhr5; GFPN hr5, pHSGFPNhr5; GFPC, pHSGFPChr5.

In the absence of AcMNPV genes, the cellular machinery is unable to promote high levels of HR

We first cotransfected pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 either alone or together into SF-21 cells and detected either no or very low levels of GFP activity (Fig. 2A, columns 1 to 3; Fig. 2C). This indicated that the cellular machinery only promoted low levels of HR and did not repair the single amino acid changes introduced into these plasmids to a functional amino acid that would yield green fluorescent cells. Similar results were obtained for the other three pairs of HR targets: pHSGFPV2TGA/pHSGFPY152TAA, pHSGFPNhr5/pHSGFPChr5, and pHSGFPN/pHSGFPC (Fig. 2B, column 1, and data not shown).

Genes necessary for AcMNPV DNA replication promote HR.

We next cotransfected pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 with AcMNPV DNA (Fig. 2, column 6; Fig. 2C) or a set of nine plasmids containing the AcMNPV replication lef genes (ie-1, ie-2, lef-1 to lef-3, lef-7, p143, dnapol, and p35) that were previously reported to be necessary for optimal viral DNA replication in SF-21 cells (24, 36) (Fig. 2A, column 7, and C). Cotransfection of the two recombination substrates with AcMNPV DNA resulted in an increase from 42 to 822 GFP-positive cells (Fig. 2A, compare columns 3 and 6). Even though the results from the transfection containing the recombination targets and replication lefs was lower than that containing AcMNPV DNA, there was a significant sixfold increase in GFP-positive cells compared to the control (Fig. 2, compare columns 3 and 7). Comparable results were obtained with the remaining GFP-containing plasmid targets (Fig. 2B and C and data not shown). It is possible that AcMNPV genes other than DNA replication genes may also stimulate HR. Transfections containing only the plasmids lacking hr5 or the plasmids lacking hr5 plus the replication lefs consistently resulted in approximately 2- to 10-fold-lower levels of HR compared to the corresponding transfections containing substrates with hr5. The addition of AcMNPV DNA to the transfection mix containing individual GFP targets resulted in only background levels of GFP-positive cells (Fig. 2A, compare columns 1 to 4 and columns 2 to 5), indicating that viral genes were not directly or indirectly repairing the mutation introduced in gfp that could have resulted in GFP activity and that a second plasmid with sequence homology was necessary for HR to occur. Similar results were obtained with the corresponding plasmid constructs lacking hr5 (data not shown).

ie-1, ie-2, lef-7, and p35 stimulate HR in the presence of hr5.

To assess the degree to which each replication lef gene contributed to HR in our assay, we omitted each replication lef systematically from the transfection mix containing all replication lefs and the reporter plasmids pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 and counted the GFP-positive cells. The removal of four individual replication lefs—lef-7, ie-1, ie-2, and p35—resulted in consistent and significant reductions to 63, 66, 59, and 52%, respectively, in GFP-positive cells (Fig. 3A, compare column 2 to columns 6, 7, 8, and 11). A minimal reduction was observed in the absence of lef-2. We decided not to investigate in detail small contributions to HR in transient assays since they could be due to variability in the assay even though they were reproducible. However, we cannot preclude that these small contributions may be essential in the context of virus infection. Interestingly, omission of lef-3 and p143 increased the levels of recombination between the DNA molecules containing gfp (Fig. 3A, compare columns 2 to 5 and 10), suggesting that these genes may be inhibitory to HR. The remaining two replication genes, lef-1 and dnapol, had no apparent effect (Fig. 3A, compare columns 2 to 3 and 9).

FIG. 3.

Requirement for specific AcMNPV replication lefs in HR. Cells were cotransfected with pHSGFPV2TGAhr5 (V2TGA hr5) and pHSGFPY152TAAhr5 (Y152TAA hr5) (A) or with pHSGFPV2TGA (V2TGA) and pHSGFPY152TAA (Y152TAA) (B) and the nine replication lefs (columns 2) or the nine replication lefs minus the lef indicated below each column (columns 3 to 12). In columns 12, the OpMNPV iap gene was included. The percentage of GFP-positive cells is reported relative to the number of GFP-positive cells produced after cotransfection with the pair of recombination substrates noted at the bottom of each panel and all of the replication lefs (A and B, columns 2, set at 100%).

We tested whether the requirement for p35 in our assay was related to its function in suppressing apoptosis. If the presence of ie-1 induced apoptosis in our assay, as previously reported (52), then p35 would be required to obtain a substantial count of GFP-positive cells. Thus, we substituted p35 with the Orgyia pseudosugata NPV inhibitor of apoptosis (Op-iap), a baculovirus gene that blocks cell death upstream of p35 (2, 37, 59). The Op-iap gene was able to substitute for p35, yielding a greater number of GFP-positive cells than were obtained with p35 (Fig. 3A, compare columns 2 to 11 and 12). If the requirement for p35 was coupled to induction of apoptosis by ie-1, we hypothesized that the withdrawal of ie-1 from the replication lef set would alleviate the need for p35. However, we were unable to test this because ie-1 was necessary for HR. Thus, the levels of HR in cells lacking both ie-1 and p35 or lacking either p35 or ie-1 were similar (Fig. 3A, compare column 13 to columns 7 and 11). An experiment directly comparing the addition of p35, ie-1, or p35 plus ie-1 with HR substrate plasmids showed that the sample containing p35 and ie-1 did not yield higher levels of HR than the one with ie-1 only (data not shown).

Major determinants for HR in the absence of hr5.

We examined HR by using plasmid targets pHSGFPV2TGA and pHSGFPY152TAA. These plasmids have the same target sequences as those used in Fig. 3A but have lower levels of DNA replication due to the absence of a cis-linked hr5 (see Fig. 6B). Target plasmids pHSGFPV2TGA and pHSGFPY152TAA were transfected alone (Fig. 3B, column 1), with the replication lefs (column 2), or lacking only one of the replication lefs (columns 3 to 11). In the absence of efficient DNA replication, we observed somewhat different requirements for HR than in the presence of higher levels of plasmid DNA replication. No reduction in HR was observed when lef-1, lef-3, or p143 was omitted from the set of replication lefs (Fig. 3B, compare column 2 to columns 3, 5, and 10). Omission of the remaining replication lefs reduced the yield of GFP-positive cells to two different levels. The omission of lef-7, ie-2, and p35 reduced the number of GFP-positive cells from 100% to ca. 40% or lower (Fig. 3B, compare column 2 to columns 6, 8, and 11), whereas the absence of lef-2, ie-1, and dnapol reduced the levels of GFP-positive cells only slightly (Fig. 3B, compare column 2 to columns 4, 7, and 9). Interestingly, we consistently observed a lower dependency for ie-1 in HR events involving targets lacking hr5 (Fig. 3).

FIG. 6.

Origin-dependent (A) or -independent (B) plasmid DNA replication. Total intracellular DNA was extracted from cells cotransfected with pHSGFPV2TGAhr5 (V2TGA hr5) and pHSGFPY152TAAhr5 (Y152TAA hr5) (A) or pHSGFPV2TGA (V2TGA) and pHSGFPY152TAA (Y152TAA) (B) alone (A, lane 2; B, lane 3) or with the complete set of replication lefs (A, lane 2; B, lane 4) or the nine replication lefs minus the gene indicated at the top of each lane. In panel A, lane 12, DNA was extracted from cells transfected with hr5-containing targets and the replication lefs except for p35 plus OpMNPV iap. GFP hr5 + B, pHSGFPhr5 was digested with BamHI to linearize the plasmid. The migration of the linear DNA is indicated at the right of each panel by an arrow. GFP hr5 + D + B, pHSGFPhr5 was digested with BamHI and DpnI to obtain the migration of DpnI-sensitive plasmid. Total cellular DNA was digested with BamHI and DpnI, and digested fragments were resolved by agarose gel electrophoresis. The DNA was transferred to a membrane and hybridized to a fragment from pHSGFPhr5 containing gfp. DNA in addition to the linear replicated plasmid that hybridized labeled gfp DNA may represent products of recombination arising from single or double-crossover events between the gfp-containing plasmids and others containing the replication lefs since these banding patterns are not as prominent in the controls (lane 1 in panel A and lane 3 in panel B), where recombination is not high and factors that stimulate recombination were not included. All of the plasmids have the hsp70 promoter and a pUC-based backbone in common.

The requirement for p35 in target plasmids lacking hr5 could be alleviated with the addition of Op-iap (Fig. 3B, compare column 2 to columns 11 and 12). Given that the requirement for ie-1 was not as pronounced for substrates lacking hr5 as it was for constructs containing hr5, we tested whether the removal of ie-1 along with p35 would demonstrate that the requirement for p35 was due to its blocking ie-1-induced apoptosis. The effects of omitting both genes were similar to those omitting ie-1 only. That is, the levels increased from just over 20% in the absence of p35 only to ca. 80% in the absence of both ie-1 and p35 (Fig. 3B, compare columns 7, 11, and 13). This implies that the involvement of p35 in HR by using the hr5-deficient substrates was to protect cells from apoptotic death induced by ie-1 expression.

Minimal requirements to promote efficient HR with substrates containing hr5.

Having defined that ie-1, ie-2, lef-7, and p35 were the major determinants stimulating HR of pHSGFPV2TGAhr5 and pHSGFPY153TAAhr5, it was important to determine whether these four genes, independent of the remaining replication lefs, were able to promote HR to levels equivalent to those with all of the replication lefs. Transfection of ie-1, ie-2, lef-7, and p35 together resulted in a greater number of GFP-positive cells than with the replication lefs (Fig. 4A, compare columns 2 and 3). The addition of individual replication lef genes to this group of four determinants of HR either increased or decreased the percentage of GFP-positive cells compared to the replication lefs (Fig. 4A, compare column 2 to columns 4 to 8). However, addition of individual replication lefs did not yield higher levels of HR than with ie-1, ie-2, lef-7, and p35 together (Fig. 4A, compare column 3 to columns 4 to 8).

FIG. 4.

Requirement for ie-1, ie-2, lef-7, and p35 in HR with substrates containing hr5. (A) Cells were cotransfected with pHSGFPV2TGAhr5 (V2TGA hr5) and pHSGFPY152TAA (Y152TAA hr5) and the nine replication lefs (column 2), or a set of four genes, ie-1, ie-2, lef-7, and p35 (column 3). In columns 4 to 8, cells were counted after cotransfection with the recombination substrates and the set of four genes listed in column 2 plus the genes indicated below each column and preceded by a plus sign (+). In columns 9 to 13, cells were scored after cotransfection with the recombination substrates and the same set of four genes minus the gene(s) listed below each column and preceded by a minus sign (−). (B) Cells were cotransfected with pHSGFPV2TGAhr5 (V2TGA hr5) and pHSGFPY152TAA (Y152TAA hr5) and the replication lefs (column 2) or a set of four genes: ie-1, ie-2, lef-7, and p35 (column 3). In columns 4 to 8, one or two genes as indicated were transfected. The percentage of GFP-positive cells is reported relative to that obtained with cells cotransfected with each pair of GFP-containing recombination substrates and the replication lefs (set to 100%, columns 2).

In order to assess the individual contribution of ie-1, ie-2, lef-7, and p35 to HR, we transfected cells with plasmids containing these four genes and systematically omitted one at a time. Surprisingly, although omission of either ie-1 or ie-2 reduced the relative levels of GFP-positive cells considerably (Fig. 4A, compare column 2 to columns 9 and 10), omission of either lef-7 or p35 did not have as great of an effect as expected (compare column 2 to columns 11 and 12). Thus, it appears that ie-1 and ie-2 are the major determinants required for HR in targets containing hr5, and lef-7 and p35 stimulate HR in the presence of ie-1 and ie-2.

Although a reaction with ie-1, ie-2, and lef-7 had ca. 90% of the relative GFP-positive cells than that with all four genes (Fig. 4A, compare columns 3 and 12), omission of p35 and ie-1 yielded levels similar to those without ie-1 only (compare columns 9 and 13), again suggesting that the role of p35 in this assay was to protect the cells from apoptosis.

In analyses defining the minimal requirements for HR, we found that ie-1 and ie-2 were the major determinants (Fig. 4A). We were then interested in determining the effects on HR that each of these genes had in our assay. The presence of ie-1 and ie-2 gave reproducibly more GFP-positive cells than the replication lefs or a sample containing ie-1, ie-2, lef-7, and p35 (Fig. 4B, compare columns 2 and 3 to 4). Independently, ie-1 or ie-2 each provided ca. 50% of the activity supplied by both genes together (compare column 4 to columns 5 and 6), suggesting independent and additive functions. In contrast, lef-7 and p35 alone did not promote HR above background levels (compare columns 1 to 7 and 8). It is possible that lef-7 and p35 are offsetting a negative effect on HR triggered by other replication lefs in Fig. 3.

Minimal requirements to promote efficient HR with substrates lacking hr5.

Figure 3B illustrates that lef-7, ie-2, and p35 make a significant contribution to HR, whereas ie-1 and dnapol may stimulate the process slightly. Since in preliminary experiments and in Fig. 5A we did not see that dnapol increased the number of GFP-positive cells in the presence of lef-7, ie-2, p35, and ie-1, we did not test its contribution in detail. However, we included ie-1 when we tested the minimal requirements to understand the effects of p35 and to compare the contribution of ie-1 to HR in the presence (Fig. 4A) or absence (Fig. 5A) of hr5.

FIG. 5.

Requirement for ie-1, ie-2, lef-7, and p35 in HR with substrates lacking hr5. Cells were cotransfected with pHSGFPV2TGA (V2TGA) and pHSGFPY152TAA (Y152TAA) and the gene(s) indicated below each column as described in the Fig. 4 legend. The percentage of GFP-positive cells is reported relative to that obtained with cells cotransfected with each pair of GFP-containing recombination substrates and the replication lefs (set to 100%, columns 2).

Supplementing ie-1, ie-2, lef-7, and p35 with each of the other replication lefs did not significantly increase or decrease the efficiency of HR (Fig. 5A, compare column 3 to columns 4 to 8). To evaluate the contribution of ie-1, ie-2, lef-7, and p35, we next omitted each gene from the transfection. A decrease in the efficiency of recombination was observed by removal of each gene (Fig. 5A, compare column 3 to columns 9 to 12). Removal of either ie-1 or lef-7 reduced the number of GFP-positive cells to 65 or 55%, respectively (compare column 3 to columns 9 and 11). This result reinforces the small effect ie-1 has in the absence of hr5. Also, in the absence of p35 and ie-1, that is, in the presence of ie-2 and lef-7 only, activity was not greater than in the absence of p35 alone (compare columns 9, 12, and 13). On the other hand, the lack of ie-2 or p35 was more substantial (columns 10 and 12) than the lack of ie-1 or lef-7 (columns 9 and 11). Whether p35 has a direct function in HR is not entirely clear at this point.

Although the effects of ie-1 in the absence of hr5 were not as pronounced as those with molecules containing hr5, we performed an experiment similar to that represented in Fig. 4B in which the contribution of each gene was tested independently and the effects of ie-1 and ie-2 were tested simultaneously. Transfection of ie-2 alone was able to provide most of the GFP activity compared to either the replication lefs or stimulation provided by ie-1, ie-2, lef-7, and p35 (Fig. 5B, compare column 6 to columns 2 and 3). On the other hand, ie-1, lef-7, or p35 alone were not able to supply considerable activity over background (compare column 1 to columns 5, 7, and 8).

HR occurs in the absence of any detectable plasmid DNA replication.

DNA replication and HR are coupled processes in a number of organisms. Nevertheless, our experiments show that HR can take place in the absence of genes required for DNA replication in transient assays (e.g., p143). Therefore, we wanted to determine to what extent our template plasmids were replicating at the time we were scoring GFP-positive cells (i.e., at 48 h posttransfection).

Cells were transfected with the recombination substrates pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 only, the substrate plasmids and all of the replication lefs, or the substrate plasmids and all of the replication lefs but one. At 48 h posttransfection, GFP-positive cells were enumerated, and DNA was extracted from cells for transient plasmid DNA replication assays. Our results were comparable to those previously published (24, 36), wherein lef-1, lef-2, lef-3, ie-1, dnapol, and p143 were essential for DNA replication, whereas lef-7, ie-2, and p35 had stimulatory effects (Fig. 6A).

A parallel experiment with target plasmids lacking hr5, pHSGFPV2TGA and pHSGFPY152TAA, confirmed that plasmid DNA replication was much less efficient without hr5 (Fig. 6B). It has been previously observed that plasmids containing early gene promoter elements are able to replicate (68). Overall, these results were similar but not identical to those obtained with plasmids containing an origin of replication in which lef-1 to lef-3, ie-1, dnapol, p143, and p35 were essential for DNA replication and lef-7 and ie-2 could be omitted without drastically affecting DNA replication.

Aphidicolin, an inhibitor of DNA polymerase α and δ, inhibits the processive elongation of DNA nascent strands but does not affect the initiation of DNA replication or the formation of short nascent strands. Aphidicolin has been widely used during baculovirus infections to block DNA synthesis and late or very late gene expression that is dependent on DNA replication. To corroborate our results that optimal DNA synthesis was not a requirement for HR, we assayed for HR in the presence or absence of aphidicolin. Cells transfected with the recombinant targets pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5, followed by infection with AcMNPV, exhibited both high levels of HR as evidenced by the number of GFP-positive cells (Fig. 7A, column 1), and efficient DNA replication, as shown by the detection of DpnI-resistant plasmid DNA in transient DNA replication assays (Fig. 7B, lane 2). Uninfected cells in the absence or presence of aphidicolin showed background levels of recombination (Fig. 7A, columns 2 and 3) and DNA replication (Fig. 7B, lanes 3 and 5). In fact, uninfected cells treated with aphidicolin showed membrane blebbing (data not shown), which is characteristic of apoptosis. However, aphidicolin-treated and AcMNPV-infected cells showed ∼7-fold more GFP-positive cells than background levels (Fig. 7A, compare columns 3 and 4) but no detectable levels of plasmid DNA replication (Fig. 7B, lane 4).

FIG. 7.

HR promoted by replication lefs in the presence of the DNA synthesis inhibitor aphidicolin. (A) Cells were transfected with pHSGFPV2TGAhr5 and pHSGFPY152TAAhr5 (columns 1 to 4). At 24 h after transfection, cells were infected with AcMNPV as indicated by the plus (+) sign (columns 2 and 4) and treated with aphidicolin after the virus adsorption period as noted by the plus sign (columns 3 and 4). Quantitation of GFP-positive cells was carried out 48 h after transfection. The relative percentage of GFP-positive cells was standardized to 100% based on cells cotransfected with recombination targets and infected with AcMNPV (column 1). (B) Total intracellular DNA was obtained from cells transfected with pHSGFPhr5 (GFP hr5, lane 1) or cotransfected with pHSGFPV2TGAhr5 (V2TGA hr5) and pHSGFPY152TAAhr5 (Y152TAA hr5) (lanes 2 to 5). Transfected cells were either infected with AcMNPV (lanes 2 and 4) or mock infected (lanes 1, 3, and 5) as indicated by the plus or minus sign, respectively. Transfected cells were also treated (lanes 4 and 5) or not treated (lanes 1 to 3) with aphidicolin as indicated by the plus or minus sign, respectively. DNA was digested, separated by agarose gel electrophoresis, transferred to a membrane, and hybridized as described in the Fig. 6 legend. The arrow on the right of the panel indicates the migration of the linear plasmids. Additional bands migrating under the DpnI-resistant linear plasmid in Fig. 6 are not prominent in this experiment since other plasmids that may recombine with the gfp substrates were not used.

DISCUSSION

We have developed an efficient and practical method to monitor and quantitate HR. Cells are transfected with reporter plasmids having homologous segments that are good substrates for recombination in the presence of genes supplied in trans that promote high-frequency HR. This transient recombination system provides a tool to both identify and assess the contribution of single or multiple genes by including or omitting them in the transfection mixture. In the present study we identified that a subset of genes involved in AcMNPV DNA replication were also required to promote high-frequency HR. This was not surprising since the processes of DNA replication and recombination have been closely linked in prokaryotes, eukaryotes, and their viruses (reviewed in reference 9). We estimate that in the presence of all of the replication lefs, the degree of HR between targets containing or lacking hr5 was 7.0 or 3.2%, respectively, within gfp in cells successfully transfected. Recombination events within other homologous regions of the substrate plasmids (e.g., hsp70 promoters, hrs, and plasmid backbones) that did not yield GFP-fluorescent cells were not estimated.

Our results show that by transfecting AcMNPV DNA, rather than the replication lef set, a higher number of GFP-fluorescent cells was observed (Fig. 2A). Four possibilities can account for this result. First, genes expressed from the genomic viral DNA containing complete 5′ and 3′ regulatory sequences may be better expressed (15) than from plasmid DNA, although we did not directly compare the levels of gene expression from these DNAs. Second, although all of the transfections in each experiment contained the same concentration of DNA, transfections where one molecule of DNA, e.g., AcMNPV DNA, was required to promote HR and not a set of DNAs encoding several genes, may be more efficient and yield higher number of GFP-positive cells. Third, the copy number of genes resulting from AcMNPV genomic DNA may be higher than that from plasmids since data were collected at 48 h posttransfection when some viral amplification may have taken place. Fourth, viral genes other than those tested stimulate HR. We are in the process of surveying the rest of the AcMNPV genome for other genes, if any, that promote efficient HR.

We also tested several other genes that have been reported to be involved in DNA replication or potentially involved in recombination events. The gene pe-38 was found to stimulate DNA replication (24), but we found no evidence of its requirement in HR (Crouch and Passarelli, unpublished). A gene with homology to the proliferating cell nuclear antigen, etl, that accelerates DNA synthesis (10, 46) also did not show any involvement in HR in our assay. Similarly, vlf-1, a gene involved in very late gene expression (42, 63) and with homologies to recombinases (42), did not stimulate HR in this assay (Crouch and Passarelli, unpublished). It is possible that VLF-1 is involved in DNA replication by resolving concatemer formation, but that function may not be apparent in this assay or a transient DNA replication assay. The conformation of plasmids that have undergone HR remains to be determined. The contribution of the AcMNPV alkaline nuclease (32) that may nick DNA generating substrates for HR has not been evaluated in this system.

HR in the presence of hr5.

Elimination of each of four genes—ie-1, ie-2, lef-7, and p35—from a DNA replication set of nine genes caused a significant decrease in HR (Fig. 3A). In the context of these four genes, ie-1 and ie-2 each provided about half of the activity and together were able to supply all or more of the activity compared to the replication lefs (Fig. 4). In contrast, lef-7 and p35 were not necessary in the context of these four genes and, individually, did not stimulate HR above background levels (Fig. 4). Thus, lef-7 and p35 stimulated HR in the presence of all of the replication lefs but were not able to promote HR in the absence of these genes or independently. It may be that lef-7 and p35 block negative effects on HR conferred by other genes of the replication apparatus. This is possible since removal of specific genes increases HR (e.g., p143 and lef-3 [see Fig. 3A]). Otherwise, lef-7 and p35 may be more directly inhibiting the HR-associated activities of ie-1 and ie-2, and these effects were obscured in the presence of other DNA replication lefs.

HR in the absence of hr5.

Cotransfections with recombination targets lacking hr5 were lower than those containing target plasmids with hr5. This finding may be related to the close association of DNA replication and recombination also seen in other systems, where DNA that is damaged during DNA replication is repaired by recombination. In addition, less-efficient DNA replication in the absence of hr5 would also result in lower number of substrates for HR.

In the presence of the eight remaining replication lefs, omission of a single DNA replication gene had an effect on HR in two tiers: HR events were more dependent on lef-7, ie-2, and p35 and less dependent on ie-1 and dnapol. A lower stimulation of HR by ie-1 in targets lacking hr5 in contrast to targets with hr5 may be attributed to the binding activity of its protein product, IE-1, to hr5 sequences. One possible scenario would be that IE-1 serves as a DNA replication initiator protein possibly recruiting recombination or replication proteins to the origin of replication. Binding of IE-1 to DNA may change the conformation of DNA, leading to more efficient HR. Binding of the Epstein-Barr virus EBNA-1 and of the herpes simplex virus type 1 UL9 to origins of DNA replication has been reported to cause DNA distortions (3, 22). With the exception of ie-1, three of the four genes necessary for recombination with plasmids containing hr5 were also major determinants in these assays, highlighting their role in HR in the presence of low or high levels of plasmid DNA replication. The activities of more lef replication genes may be necessary in the absence of hr5 to either initiate or proceed with DNA synthesis where either enough single strands of DNA or DNA lesions are available that can be substrates for recombination. However, supplying dnapol to ie-1, ie-2, lef-7, and p35 did not stimulate HR. As in the case with substrates containing hr5, lef-7 and p35 individually did not promote HR above background levels (Fig. 5B). Additionally, ie-1 alone had a minimal effect on HR. Surprisingly, over half of the activity was contributed by ie-2 alone. Thus, it appears that ie-2 is the major effector of HR in the absence of hr5. This may be related to the ability of ie-2 to induce cell cycle arrest (53). Whether the same genes are required for HR during the recombination of viral genomes remains to be determined.

Role of p35 in HR.

The role of p35 in HR can be attributed to its ability to protect cells from undergoing apoptosis induced by either ie-1, which has been previously shown to stimulate apoptosis (52), or some other stimulus. Op-iap, another suppressor of apoptosis that works at a different point in the apoptotic pathway, effectively replaced p35. When we tested whether the requirement for p35 was to block cell death induced by ie-1, we found that the lack of p35 and ie-1 using recombination substrates with hr5 did not result in a lower number of GFP-positive cells than reactions lacking ie-1 only. These data suggest that the role of p35 is to block cell death.

In experiments with targets lacking hr5 in which ie-1 was not as strongly involved in HR, we found that p35 still had no effect on HR in the absence of ie-1 when tested in the presence of the remaining replication lefs but that it had a slight effect in the presence of lef-7 and ie-2 only. We thought that this small effect on HR by p35 could result whether either ie-2 or lef-7 induced apoptosis. However, neither lef-7 nor ie-2 appears to induce apoptosis in transfections (53; Crouch and Passarelli, unpublished). Although at this time we favor the possibility that the involvement of p35 in HR is to prevent apoptosis, it is possible that p35 has some role on HR or DNA replication in vivo. It has been observed that higher levels of p35 are necessary to support very late gene expression than levels required to prevent apoptosis (21). Adequate levels of very late gene expression are reflective of successful DNA replication.

HR in the absence of optimal DNA replication.

Results from our plasmid DNA replication experiments parallel previously published data. Like Lu and Miller (36), we found that lef-7 stimulated DNA replication in contrast to the findings of Kool et al. (24), who found that it was not required. However, our results concerning dnapol and p35 corroborate those of Kool et al. We found that dnapol and p35 were essential and stimulatory for DNA synthesis, respectively, whereas Lu and Miller found that dnapol was stimulatory and p35 was essential. A number of explanations could account for these slight discrepancies. First, although we are using the same hr5 region as in the Lu and Miller study, our set of replication lefs and reporter plasmids are different. In both of the published studies the lef genes were under their native promoter control and may have been expressed at different levels. We used the hsp70 promoter, and the genes should be more uniformly expressed and in some cases were probably overexpressed. Second, we extracted DNA from transfected cells at 48 h rather than at 72 or 96 h posttransfection as in the previous studies. It was suggested previously that a delay in plasmid DNA replication in the absence of dnapol could appear more pronounced at an earlier time (36). In addition, the dependence of p35 at earlier times, when apoptosis is not as prominent and/or when the apoptotic signal is still being mounted, has been suggested or demonstrated elsewhere (63; Crouch and Passarelli, unpublished).

We could not detect plasmid DNA replication in the absence of p143. Thus, in the absence of detectable DNA replication, HR still took place. This result was puzzling at first since DNA replication and recombination have been linked. We found that we required genes that have a function in DNA replication for HR, stressing the concurrency of these two mechanisms. To test whether HR was independent of DNA replication, we performed another experiment with the DNA synthesis inhibitor aphidicolin. We observed only a 20% decrease in GFP-fluorescent cells in infected cells treated with aphidicolin (Fig. 7A). The absence of complete and detectable plasmid DNA replication in the assay was confirmed by analyzing DpnI sensitivity in the input plasmids. Aphidicolin inhibits viral DNA polymerase (44) and cellular DNA polymerases α and δ. Since aphidicolin inhibits the processive elongation of DNA strands but not the initiation of DNA replication or the production of short nascent strands, we cannot rule out the possibility that HR is dependent on the initiation of DNA replication or host factors, including the host DNA polymerase β, that is not inhibited by aphidicolin and is involved in DNA repair.

Martin and Weber (38) conducted a study that used inversion of inverted repeat sequences either inserted in AcMNPV or as part of a plasmid in Sf9 cells as an indication of recombination. Recombination in that study was strictly dependent on the presence of hr2, and 7 out of the 9 replication lefs (p143, lef-1 to lef-3, ie-2, dnapol, and p35) expressed under native promoter control. Although our results appear to conflict with those of Martin and Weber, there are a number of differences between the two systems, including the hr used, the promoters used to express the replication lefs, the time at which recombination was analyzed, and the cell line used. Perhaps most importantly, Martin and Weber monitored intramolecular recombination, whereas we examined intermolecular events. We were more interested in the requirement for intermolecular recombination since this is the type of recombination that occurs between baculovirus genomes or during construction of baculovirus expression vectors.

Mechanism of HR in AcMNPV.

Interestingly, three genes that were found to stimulate but were not essential for DNA replication are involved in HR. One possible model in a transient DNA replication assay would be that a set of genes initiates and allows elongation of nascent DNA strands. A second set of genes, the DNA replication stimulatory genes, renders more efficient DNA replication by allowing elongation to proceed. Alternatively, it can be envisioned that once DNA replication is initiated it reaches blocks that stall the replication fork, and genes that stimulate DNA replication allow the process to continue since these genes repair the defects (i.e., DNA bound by proteins or DNA breaks) via HR. It is thought that the primary function of HR is the repair of halted or collapsed replication forks to maintain genome integrity in both eukaryotes and prokaryotes (reviewed in reference 9).

Although we found that ie-2 was the main contributor for HR, we have no evidence that it serves as a viral recombinase. It is clear from Fig. 3A that no single gene, including ie-2, reduces the levels of HR to background levels, implying that the remaining viral genes or the viral genes together with the host genes are providing low levels of HR. Although ie-2 has been studied for a number of years as a transactivator of early, late, and very late genes in the presence of ie-1 or other genes (5, 6, 48), very little is known about how it works. The requirement for ie-2 in both late gene expression and plasmid DNA replication is host specific (35). More recently, it was shown that ie-2 blocks cell cycle progression from the S phase in cell lines derived from several species (53). One possibility is that IE-2 is arresting the cell cycle to allow viral DNA replication and repair. We are currently investigating the specific role of ie-2 in recombination and its role in the cell cycle by using several cell lines.