A Conserved N-Terminal Domain Mediates Required DNA Replication Activities and Phosphorylation of the Transcriptional Activator IE1 of Autographa californica Multicapsid Nucleopolyhedrovirus

Abstract

IE1 is the principal transcriptional regulator of the baculoviruses. Like multifunctional transcription factors of other large DNA viruses, IE1 is an essential, site-specific DNA-binding phosphoprotein that activates virus gene expression and promotes genome replication. To define the poorly understood mechanisms by which IE1 achieves its diverse functions, we identified IE1 domains that contribute to productive infection of Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV), the baculovirus prototype. Site-directed mutagenesis revealed that the N-terminal 23 residues of IE1 are required for origin-specific DNA replication and AcMNPV propagation, but not for DNA-binding-dependent transcriptional activation. Within this defined replication domain, we identified an invariant TPXR/H motif that resembles a consensus cyclin-dependent kinase phosphorylation site. Amino acid substitutions of potential phosphorylation sites within or near this motif caused loss of IE1-mediated DNA replication activity. Remarkably, substitution of the single threonine (residue 15) within the TPXR/H motif caused complete loss of AcMNPV multiplication. The replication domain was required for IE1 phosphorylation. It was also sufficient for conferring phosphorylation of a heterologous protein. Importantly, IE1 hyperphosphorylation coincided exclusively with AcMNPV DNA replication. The temporal regulation of IE1 phosphorylation and the essential nature of the TPXR/H motif suggest that phosphorylation critically alters and possibly activates DNA replication activity of IE1 during infection. The striking conservation of the TPXR/H motif among IE1 proteins further suggests that this molecular switch may be a common mechanism by which the alphabaculoviruses coordinate DNA replication and gene expression by using a single regulator.

INTRODUCTION

The baculoviruses are a family of large DNA viruses that are prolific insect pathogens and highly efficient eukaryotic expression vectors. The prototypic species is Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV). Like many DNA viruses, the coordination of gene expression and genome replication are critical for baculovirus multiplication (reviewed in references 6 and 39). The principal transcriptional regulator of AcMNPV is the immediate-early protein IE1, a 582-residue DNA-binding phosphoprotein that is essential for virus multiplication (43, 46). IE1 activates early and late viral gene transcription and also promotes origin-specific DNA replication. How IE1 accomplishes these diverse functions is poorly understood. Because multifunctional transcriptional regulators are essential for genome replication of many large DNA viruses, the molecular mechanisms of activators like IE1 are of great interest.

Transactivators of DNA viruses often have specific domains responsible for different activities. Distinct functional domains define the organizational structure of AcMNPV IE1. The N-terminal-third of IE1 mediates transcriptional activation (Fig. 1). This separable section confers a >200-fold increase in transcription from select AcMNPV promoters (4, 18, 38, 44). The relatively acidic transactivation domain is divided by a small basic residue stretch called basic domain I (BDI) (Fig. 1), which is required for IE1 sequence-specific DNA binding (31). At its C terminus, IE1 contains another basic stretch (basic domain II) and a helix-loop-helix dimerization motif which are required for DNA binding and nuclear import (29, 30). IE1 binds as a dimer to the 28-bp imperfect palindromic repeats (28-mers) that constitute the transcriptional enhancer activity of the homologous regions (hrs), which are distributed throughout the circular DNA genome (134 kb) of AcMNPV (9, 37). Sequence-specific binding to these 28-mer repeats is necessary for enhancer-dependent transcriptional stimulation by IE1 (7, 9, 36, 37).

Fig 1.

Functional domains of AcMNPV IE1. (A) Structure of full-length IE1. The 582-residue AcMNPV IE1 protein (AcIE1) possesses a transactivation domain (residues 1 to 222), a sequence-specific DNA-binding domain (residues 152 to 161) designated basic domain I, an unrelated basic domain (residues 534 to 538) designated basic domain II required for nuclear import, and a C-terminal oligomerization domain (crosshatched) required for dimerization. (B) IE1 deletions. Each constructed deletion is designated (left) according to the residues removed from within the N-terminal half of AcIE1, which contains the transactivation domain (open box). (C) The AcIE1 N terminus. Residues comprising the replication domain (see below), the transactivation domain, and BDI are indicated; acidic (−) or basic (+) residues are marked. The AcIE1 residues encompassed by each deletion tested here are underlined.

IE1 is an origin-binding protein by virtue of its capacity to interact with the hrs, which function as origins of viral DNA replication (7–9, 21, 33, 36, 37). Consistent with its replicative activity, IE1 colocalizes with viral DNA and essential viral DNA synthesis proteins within nuclear replication centers of infected cells (15, 25, 28). When tested in transfected cells, IE1 participates in combination with virus-encoded DNA helicase P143, primase factors LEF-1 and LEF-2, single-stranded DNA-binding protein LEF-3, and DNA polymerase (DNA Pol) to replicate hr-containing DNA plasmids (17, 35). This DNA replication activity is independent of IE1's transcriptional activity (35). Although IE1 is an essential DNA replication factor, the mechanism by which it participates in virus DNA synthesis is unknown.

To determine the role of IE1 in viral DNA replication and to better define the poorly understood mechanisms of baculovirus genome synthesis, we first sought to identify IE1 domains that are essential for DNA replication and virus multiplication in permissive Spodoptera frugiperda cells. Previous studies of the alphabaculovirus Orgyia pseudotsugata MNPV (OpMNPV) revealed that the N-terminal 65 residues of OpMNPV IE1 are required for plasmid DNA replication assays in transfected cells (32). However, this N-terminal region and the complete transactivation domain of OpMNPV IE1 exhibit little if any sequence identity with AcMNPV IE1. Consequently, we scanned AcMNPV IE1 for residues critical in DNA replication. We report here that IE1's N terminus contains a 23-residue replication domain that is required for virus DNA synthesis and virus multiplication but is dispensable for IE1 transcriptional activation. This replication domain also contributes to phosphorylation of IE1 in a process that is temporally regulated to coincide with genome replication during infection. We subsequently discovered a cyclin-dependent kinase consensus motif (TPXR/H) within the replication domain that is critical for AcMNPV multiplication. Our study suggests that phosphorylation plays an essential regulatory role in the DNA replication activity of IE1. Moreover, the striking conservation of the TPXR/H motif at the N terminus of IE1 proteins suggests that a common mechanism is used among alphabaculoviruses to temporally regulate IE1's replicative and transcriptional activities for optimum virus multiplication.

MATERIALS AND METHODS

Plasmids. (i) IE1^HA and IE1^FLAG.

Expression plasmid pIE1^HA/_BS carries wild-type AcMNPV IE1 (AcIE1) with a hemagglutinin (HA) epitope tag (residues YPYDVPDY) inserted after C-terminal IE1 residue 579 (30) and placed under the control of the native ie-1 gene promoter. The NdeI-SstI fragment of pSP64/AcIE1^FLAG (30) was inserted into pAcIE1^HA/_BS to generate pAcIE1^FLAG/_BS, which carries an AcMNPV IE1 epitope tagged with FLAG (residues DYKDDDDK) at its C terminus.

(ii) IE1^HA deletions.

To generate expression vectors encoding deletions of AcMNPV IE1^HA, specific IE1 residues were replaced with a Gly-Ser insertion. To this end, 5′ sections of ie-1, including its promoter and codon(s) 1, 1 to 7, 1 to 15, 1 to 23, 1 to 32, 1 to 64, 1 to 89, 1 to 124, and 1 to 136 were PCR amplified using primers containing HindIII and BamHI sites. Primer sequences are available upon request. Corresponding sections of ie-1, including codons 2 to 430, 8 to 430, 16 to 430, 24 to 430, 33 to 430, 65 to 430, 90 to 430, 125 to 430, and 137 to 430 were amplified with primers containing BamHI sites. The PCR products containing 5′ and 3′ sections were digested with HindIII/BamHI or BamHI/NdeI, respectively, ligated, and inserted into HindIII- and NdeI-digested pIE1^HA/_BS to generate vectors pAcIE1(Δ2–32)-HA, pAcIE1(Δ3–64)-HA, pAcIE1(Δ65–89)-HA, pAcIE1(Δ90–124)-HA, pAcIE1(Δ125–136)-HA, pAcIE1(Δ2–7)-HA, pAcIE1(Δ8–15)-HA, pAcIE1(Δ16–23)-HA, and pAcIE1(Δ24–32)-HA. All plasmids and their alterations were verified by nucleotide sequencing.

(iii) IE1^HA substitutions.

The IE1 substitutions T2A, S9A/Y10F, T11A/S12A, S14A/T15A, T15A, S17A/S20A, and S24A/Y25F/S26A were generated by site-directed mutagenesis of pIE1^HA/_BS by using standard PCR methods.

(iv) IE1-GFP^HA fusions.

Plasmid pGFP^HA encodes a C-terminally HA epitope-tagged enhanced green fluorescent protein (GFP) expressed from the AcMNPV ie-1 promoter. To generate pGFP^HA, sequences encoding residues 2 to 260 of GFP were PCR amplified from pIE1^hr/EGFP-IE1/PA (29) using primers containing BamHI and AatII sites, digested with BamHI/AatII, and inserted into the corresponding sites of pAcIE1(Δ2–32)-HA. The pAcIE1(1–7)-GFP^HA, pAcIE1(1–15)-GFP^HA, pAcIE1(1–32)-GFP^HA, and pAcIE1(1–64)-GFP^HA expression vectors were generated by inserting the HindIII-BamHI fragments of vectors pAcIE1(Δ8–15)-HA, pAcIE1(Δ16–23)-HA, pAcIE1(Δ33–64)-HA, and pAcIE1(Δ65–89)-HA, respectively, into pGFP^HA.

(v) Luciferase reporters.

Reporter plasmids pFL35K-Luc/PA, containing the luciferase gene (luc) under the control of the full-length p35 promoter (−110 to +15), and pBAS35K-Luc/28mer-up+/PA, containing luc under the control of the p35 basal promoter (−30 to +15) cis-linked to a single 28-mer enhancer element of the AcMNPV hr5 enhancer, were described previously (30).

(vi) Replication factor expression vectors.

A library of late expression factor (lef) expression plasmids (pHSEpiHis) that carry N-terminal HA-tagged AcMNPV replication factors under the control of the Drosophila melanogaster heat shock protein 70 promoter (35) were kindly provided by Lorena Passarelli (Kansas State University). Plasmid pIE1^prm-P35, carrying untagged P35 under the control of the ie-1 promoter, was generated by deleting the hr5-containing BamHI-BamHI fragment from plasmid pIE1^hr-P35 (11).

Cells.

Spodoptera frugiperda IPLB-SF21 cells (50) and Drosophila melanogaster DL-1 cells (41) were maintained at 27°C in TC100 (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS; HyClone) or Schneider's growth medium (Invitrogen)–15% FBS, respectively. Stably transfected SF21 cells were generated as described previously (2). After transfection with pIE1^FLAG/_BS and pIE1/neo/PA, carrying the gene for neomycin resistance, cells were selected for resistance to Geneticin (G418 sulfate; Gibco) and pooled; drug-resistant cell lines were propagated in the absence of G418. Constitutive production of IE1^FLAG was confirmed by immunoblot analysis. The stable IE1^FLAG cells were used to titer ie-1-deficient AcMNPV by end point dilution assays, which provided 50% tissue culture infective dose (TCID₅₀) values.

Viruses.

Wild-type (wt) L-1 strain AcMNPV and AcMNPV recombinants wt/lacZ (p35⁺ iap⁻ polyhedrin gene [polh]¯ lacZ⁺) were described previously (10, 20). For inoculation, extracellular budded virus in TC100–10% FBS was added to SF21 or DL-1 monolayers at a multiplicity of infection (MOI) of 10 PFU per cell. After 1 h at room temperature, the inoculum was replaced with serum-supplemented medium, and cells were incubated at 27°C. Where indicated, the cells were overlaid 3 h after inoculation with supplemented medium containing 10 mM caffeine (Alexis Biochemicals).

AcMNPV bacmid rescue assay.

The ie-1 knockout bacmid AcBIE1KO (46), kindly provided by David Theilmann (University of British Columbia), was maintained in Escherichia coli strain DH10B. In AcBIE1KO, a zeocin resistance gene replaces the ie-1 open reading frame within AcMNPV bacmid bMON14272 (Invitrogen Life Technologies). In addition, a GFP reporter under the control of the early OpMNPV ie-1 promoter is inserted adjacent to the polyhedrin gene under the control of its own promoter (46). In the complementation assay, SF21 monolayers (10⁶ cells per 60-mm plate) were transfected with purified AcBacIE1KO DNA (1 μg) either alone or with the indicated plasmid carrying the IE1^HA gene (5 μg). After 72 h, cells and growth medium were collected. The cells were lysed in 1% sodium dodecyl sulfate (SDS) and 1% β-mercaptoethanol (βME) and subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and immunoblot analysis. Stable SF21^IE1 cells that constitutively produced IE1^FLAG were used along with the method of end point dilution (27) to measure titers of infectious budded virus in the growth medium.

Transfections.

SF21 monolayers were overlaid with TC100 medium containing CsCl-purified plasmid DNA and cationic liposomes consisting of N-[1-(2,3 dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl sulfate-l-α-phosphatidylethanolamine, dioleoyl (C18:1; [cis]-9; DOTAP-DOPE) as described previously (19). After 4 h at room temperature, the transfection mix was replaced with supplemented TC100. When indicated, the phosphatase inhibitor okadaic acid (Sigma) was dissolved in dimethyl sulfoxide (DMSO), diluted in supplemented TC100 to the indicated concentration (nanomolar levels), and used to overlay plasmid-transfected monolayers (10⁶ cells per 60-mm plate) for 2 h at 27°C.

Luciferase reporter assays.

SF21 cells (10⁶ cells per 60-mm plate) were transfected in duplicate with luciferase reporter plasmid (1 μg) either alone or with the indicated IE1^HA-encoding vector (1 μg). Cells were collected 24 h later, washed with phosphate-buffered saline (20), and lysed by suspension in 1× cell culture lysis reagent (Promega). After clarification by centrifugation, luciferase activity from 20 μl of cell lysate was measured by using a luminometer (Monolight 3010).

Transient DNA replication assays.

SF21 cells (10⁶ cells per 60-mm plate) were transfected with pBas35K-Luc/Achr5/PA template plasmid (1 μg) containing the AcMNPV hr5 origin of replication, plasmid (1 μg) encoding the indicated IE1^HA P35 expression plasmid (0.5 μg) pIE1^prm-P35 and pHSEpiHis LEF expression plasmids (0.5 μg each) carrying genes for LEF-1, LEF-2, LEF-3, P143, DNA polymerase, LEF-7, and IE2. Cells were harvested 48 h later and suspended in 10 mM Tris (pH 8.0), 1 mM EDTA, 0.2% SDS, and 1 mg/ml proteinase K. After 4 h at 37°C, the mixture was phenol-chloroform extracted and treated with 100 μg of RNase A/ml. Nucleic acid was precipitated with ethanol and suspended in 10 mM Tris (pH 8.0)–1 mM EDTA. DNA from the equivalent of 10⁵ cells was digested at 37°C for 12 h with 8 U DpnI to remove the original (unreplicated) template plasmid and 8 U XbaI to linearize the newly replicated template plasmid. Digested DNA from the equivalent of 2.5 × 10⁴ cells was subjected to agarose gel electrophoresis, transferred to Nytran membrane (Whatman), and hybridized to an XbaI-EcoRV fragment of pBas35K-Luc/Achr5/PA containing the luciferase gene radiolabeled with [³²P]dCTP (random primer labeling kit; Invitrogen). Signals were visualized by PhosphorImager analysis. Signal intensity was quantified by densitometry (ImageJ; NIH), and images were prepared by using Adobe Photoshop CS4 and Illustrator CS4.

Immunoblot assays and antisera.

Cells were collected by centrifugation, lysed in 1% SDS–1% βME, and subjected to SDS-PAGE. Proteins were transferred to nitrocellulose (Osmonics, Inc.) or Immobilon-P polyvinylidene difluoride (Millipore) membranes and detected with antisera diluted as indicated: anti-HA (1:2,000; Covance), polyclonal anti-IE1 (1:10,000 [30]), anti-actin (1:5,000; BD Bioscience), or monoclonal anti-P35 (1:10,000; kindly provided by Yuri Lazebnik, Cold Spring Harbor Laboratory). Signal development was conducted by using alkaline phosphatase-conjugated goat anti-rabbit or goat anti-mouse immunoglobulin G (Jackson ImmunoResearch Laboratories, Inc.) with the CDP-Star chemiluminescence detection system (Roche Diagnostics). Films were digitized with an Epson Twain Pro scanner and prepared by using Adobe Photoshop CS4 and Illustrator CS4.

Phosphatase treatment.

Plasmid-transfected SF21 cells were collected, freeze-thawed using a dry ice-ethanol bath, and suspended in phosphatase buffer (50 mM Tris-HCl [pH 7.9], 10 mM MgCl₂, 100 mM NaCl, 1 mM dithiothreitol) with 1× Complete protease inhibitor (Roche). After clarification by centrifugation, the supernatant was divided into three equal fractions. One fraction was diluted immediately to give 1% SDS–1% βME and heated for 2 min at 95°C to inhibit intracellular phosphatase. The other two fractions were treated for 1 h at 37°C with 5 U of calf intestinal phosphatase (CIP; Promega) or 50 mM EDTA, respectively. These samples were then treated as the first samples, and all samples were subjected to high-resolution SDS-PAGE and immunoblot analysis.

RESULTS

N-terminal IE1 residues are required for productive infection by AcMNPV.

IE1 is a sequence-specific DNA-binding protein that functions as a potent transcriptional activator. Yet, the exact functions of this essential immediate-early protein during the alphabaculovirus infection cycle are poorly understood. To define in vivo activities of IE1, we developed a sensitive complementation assay that assessed the effect of site-specific mutations within IE1 on multiplication of AcMNPV. The assay tests the ability of wild-type IE1 or mutations thereof to rescue replication of an AcMNPV ie-1 knockout (KO) bacmid when introduced into permissive SF21 cells by transfection. The KO bacmid is replication defective because it lacks the ie-1 gene and its overlapping splice variant, ie-0 (46). It also encodes a constitutive GFP reporter, which allows direct quantitation of transfection efficiency.

Confirming the absolute requirement for IE1 (46), the ie-1 KO bacmid failed to produce infectious progeny, as indicated by a minor accumulation of GFP-expressing cells from 24 h to 72 h after transfection (Fig. 2A), and undetectable levels of budded virus at late times after transfection (Fig. 2B). In contrast, when a plasmid carrying the gene for wild-type IE1^HA was transfected with the ie-1 KO bacmid, virus multiplication was restored, as indicated by widespread GFP expression (Fig. 2A) and high yields of budded virus, ranging from 10⁵ to 10⁶ PFU per ml of culture medium (Fig. 2B). Thus, plasmid-expressed wild-type IE1^HA complemented the ie-1 KO bacmid. We used this assay to identify IE1 domains required for function. Previous studies (18, 38, 44) mapped N-terminal residues involved in IE1 transcriptional activation (Fig. 1A). To investigate the role(s) of these residues during infection, we first tested the effects of deletions within the transactivation domain (Fig. 1B) on the capacity of IE1 to complement the ie-1 KO bacmid. BDI divides the transactivation domain (Fig. 1) and is required for IE1 binding to the AcMNPV hr enhancer elements (31). Because IE1 residues C-terminal to BDI contribute to DNA binding (18, 22, 31, 38), they were not analyzed here.

Fig 2.

N-terminal IE1 residues required for AcMNPV multiplication. (A) Live cell images. SF21 cells were transfected with ie-1 knockout bacmid AcBacIE1KO (1 μg) alone or with plasmid (5 μg) carrying the gene for wt IE1^HA or deletions thereof. Bright-field photographs and fluorescence images were taken 24 and 72 h after transfection to display the fraction of cells transfected successfully with the bacmid carrying the GFP reporter (24 h) and subsequently infected by newly produced GFP-encoding budded virus (72 h). Representative images (magnification, ×100 magnification) are shown. (B) Virus yields. Extracellular budded virus produced by SF21 cells transfected with AcBacIE1KO and the indicated IE1^HA-carrying plasmids as described in panel A was quantified 72 h after transfection by TCID₅₀ measurements using SF21 cells that stably produced IE1^FLAG to complement any ie-1 KO viruses produced. Reported values are averages ± standard deviations of virus yields (in PFU per ml) from three independent transfections. (C) Immunoblots. Whole-cell lysates prepared 72 h after transfection of SF21 cells as described in panel A were subjected to immunoblot analysis by using anti-HA (top), anti-P35 (middle), or anti-actin (bottom) to demonstrate protein loading levels. Size standards (in kilodaltons) are indicated on the right.

When the ie-1 KO bacmid was transfected with plasmids carrying genes for internal deletions of the transactivation domain, including IE1(Δ33–64)^HA, IE1(Δ65–89)^HA, IE1(Δ90–124)^HA, and IE1(Δ125–136)^HA, virus yields declined from 10- to 1,000-fold compared to wild-type IE1^HA (Fig. 2B). The transfection efficiency was comparable for each assay as judged by quantification of virus-mediated GFP fluorescence 24 h after transfection (data not shown). We concluded that transactivation is a critical IE1 function, but that no one internal section of the transactivation domain (Fig. 1C) is absolutely required. In contrast, deletion IE1(Δ2–32)^HA, which lacks the first 32 residues of IE1, failed to support virus multiplication (Fig. 2B). Thus, the N-terminal 32 residues contribute to an essential function of IE1.

The steady-state level of each IE1 deletion in the complementation assay was comparable to that of wild-type IE1^HA (Fig. 2C, lanes 2 to 7). Thus, the loss of function of IE1(Δ2–32)^HA was not due to protein instability. To qualitatively assess transcriptional activation of the IE1 deletions during complementation, we monitored expression of the bacmid carrying the gene for p35. Early genes like p35 are dependent upon IE1-mediated transactivation for expression (43, 46). This dependence was demonstrated by the inability to detect P35 in cells transfected with ie-1 KO bacmid alone (Fig. 2C, lane 1). Conversely, P35 was abundant upon complementation with wild-type IE1^HA and each of the IE1 deletions (lanes 2 to 7). However, P35 levels were lowest when complemented by N-terminal deletion IE1(Δ2–32)^HA. This finding suggested that either IE1(Δ2–32)^HA was partially defective for transactivation or that the p35 dosage was reduced due to defective genome replication that limited cell-to-cell spread of virus over the 72-h period.

The IE1 N terminus is dispensable for transactivation.

To more accurately evaluate transcriptional activation by the IE1 deletions, we used transient-expression reporters in the absence of virus to avoid variations due to gene dosage. To this end, we used a luciferase gene reporter driven by the basal promoter (TATA box and RNA start site) of the well-characterized p35 promoter, which was linked to either a single 28-mer enhancer repeat from the AcMNPV hr5 element (hr-dependent reporter) or the p35 upstream activating region (UAR; UAR-dependent reporter) (Fig. 3A). In combination with these cis-acting signals, the p35 promoter is highly sensitive to IE1 transactivation (26, 30, 37). Immunoblot assays demonstrated that levels of wild-type or deleted IE1^HA were comparable upon transfection of expression plasmids and the reporter (Fig. 3B). Compared to reporter activities in the absence of IE1, hr- and UAR-dependent reporter expression increased by >1,500- and 250-fold, respectively, in the presence of wild-type IE1^HA (Fig. 3C). Transactivation of the hr-dependent reporter conferred by the IE1 deletions varied from 10% to 60% of that by wild-type IE1^HA (Fig. 3C). Transactivation of the UAR-dependent reporter paralleled that of the hr-dependent reporter. In particular, transactivation by the multiplication-defective deletion IE1(Δ2–32)^HA was comparable to that of the replication-competent internal deletions IE1(Δ33–64)^HA, IE1(Δ90–124)^HA, and IE1(Δ125–136)^HA. This finding confirmed that IE1's N-terminal 32 residues are not required for IE1-mediated transactivation. Rather, these residues must be involved in a different multiplicative function.

Fig 3.

Transcriptional activation by IE1 deletions. (A) hr- and UAR-dependent reporters. The luciferase reporter gene was placed under the control of the p35 basal promoter, consisting of the TATA element and RNA start site (+1), cis-linked to a single 28-mer enhancer element from AcMNPV hr5 or the full-length AcMNPV p35 promoter with its UAR from nucleotides −110 to −30. (B) Immunoblots. SF21 cells were transfected with reporter plasmid (1 μg) alone (none) or with plasmid (1 μg) carrying the gene for wt IE1^HA or the indicated IE1^HA deletions. Total cell lysates prepared 24 h after transfection were subjected to immunoblot analysis by using anti-HA; a typical immunoblot is shown. Size standards (in kilodaltons) are indicated. IE1^HA accumulation was comparable in the presence of either reporter plasmid. (C) Reporter assays. Extracts from cells transfected with plasmids 24 h earlier as described in panel B were assayed for active luciferase. The values reported are the average luciferase activities ± standard deviations of the p35 UAR-containing (crosshatched) or Ac28-mer-containing (solid) reporters obtained from triplicate transfections and normalized to that obtained for cells transfected with plasmid carrying wild-type IE1^HA. Luciferase activities from the hr- and UAR-dependent reporters alone (none) increased >1,500- and 250-fold, respectively, when plasmid carrying wt IE1^HA was included.

N-terminal residues of IE1 are required for DNA replication.

In addition to transcriptional activation of baculovirus genes, IE1 may have a direct role in viral DNA synthesis, as first suggested from transient DNA replication assays (17, 35). To further define the replicative functions of IE1, we tested our IE1 deletions for their capacity to support plasmid DNA replication in the presence of constitutively expressed AcMNPV replication factors. To this end, SF21 cells were transfected with an hr5 origin-containing template plasmid, expression vectors for constitutive production of eight viral replication factors (LEF-1, -2, and -3, P143, DNA Pol, IE2, P35, and LEF-7), and plasmids for expression of each of the IE1 deletions. The Drosophila heat shock protein 70 promoter was used to drive expression of the required virus-derived replication factors (35); thus, their expression was independent of any loss of transcriptional activation by potentially defective IE1 mutations included in the assay. Production of each of the IE1^HA deletions and HA-tagged viral replication factors was confirmed by immunoblot analysis (data not shown). The hr5-containing template plasmid was not replicated in the absence of IE1 (Fig. 4A, lane 2). However, wild-type IE1^HA promoted hr5 template DNA replication, as indicated by detection of a DpnI-resistant signal representing newly synthesized template plasmid (lane 3). Likewise, each of the IE1^HA internal deletions that supported virus multiplication also supported template replication (lanes 5 to 8). Quantitation of triplicate assays (Fig. 4B) revealed that IE1(Δ33–64)^HA, IE1(Δ65–89)^HA, IE1(Δ90–124)^HA, and IE1(Δ124–136)^HA supported template replication at levels equivalent to or greater than that of wild-type IE1. In contrast, the N-terminal deletion IE1(Δ2–32)^HA failed to support replication (Fig. 4A and B). We concluded that IE1 possesses a domain between residues 1 and 32 required for IE1-mediated DNA replication. In contrast, transactivation residues 33 to 136 were dispensable for the same function. It should be noted that these internal residues might contribute to viral genome replication indirectly by affecting IE1-mediated transactivation of AcMNPV replicative genes.

Fig 4.

hr5-dependent DNA replication by IE1 deletions. (A) Replicated plasmid DNA. SF21 cells were cotransfected with an hr5-containing template plasmid (1 μg), plasmids (0.5 μg) carrying genes for each of the required AcMNPV replication factors (LEF-1, -2, -3, and -7, helicase P143, DNA Pol, IE2, and P35), and plasmid (1 μg) carrying the gene for wild-type (wt) IE1^HA or the indicated IE1 deletion or no IE1 plasmid (none). Total intracellular DNA isolated 48 h after transfection was digested with XbaI and DpnI. Replicated template plasmid was quantified by Southern blot hybridization to a ³²P-labeled DNA probe complementary to the hr5 of the template plasmid; a representative blot is shown. For the appropriate size marker, XbaI-linearized hr5-containing (8,950-bp) plasmid (template) was included (lane 1). (B) Quantitation. Values obtained by PhosphorImager analysis of triplicate blots are reported as the relative averages ± standard deviations of replicated DNA obtained from triplicate transfections and normalized to that of cells transfected with wild-type IE1^HA-carrying plasmid.

IE1 residues 2 to 23 constitute a replication domain.

To further define the function of N-terminal IE1 residues, we generated another series of IE1 deletions between residues 2 and 32. Wild-type IE1^HA and deletion IE1(Δ24–32)^HA supported comparable levels of budded virus production when tested in the ie-1 KO bacmid assay (Fig. 2B). In contrast, IE1(Δ2–7)^HA, IE1(Δ8–15)^HA, and IE1(Δ16–23)^HA failed to complement the bacmid. Immunoblot analysis confirmed that each of the IE1 deletions accumulated to similar steady-state levels (Fig. 2C, lanes 8 to 11), indicating that the loss of function of these deletions was not due to protein instability. Each of these IE1 deletions also stimulated IE1-dependent P35 production, suggesting that each mutation was transactivation competent. Indeed, transactivation of luciferase reporters by each of the IE1 deletions was reduced by no more than 20% compared to that of wild-type IE1 (Fig. 3C). Moreover, transactivation by replication-defective IE1(Δ2–7)^HA, IE1(Δ6–15)^HA, or IE1(Δ16–23)^HA was greater than that of replication-competent IE1(Δ33–64)^HA, IE1(Δ65–89)^HA, IE1(Δ90–124)^HA, or IE1(Δ124–136)^HA. Thus, the loss of function of the replication-defective deletions was not due to a lack of IE1-stimulated gene expression. As judged by transient DNA replication assays, deletions IE1(Δ2–7)^HA, IE1(Δ8–15)^HA, and IE1(Δ16–23)^HA failed to support replication of an hr5-containing template plasmid (data not shown). We concluded that IE1 residues 2 to 23 constitute a domain, here referred to as the replication domain, which is required for a replicative function of IE1.

The replication domain is required for IE1 phosphorylation.

IE1 is a nuclear phosphoprotein (3, 44), but the biochemical role of phosphorylation and the specific residues involved are unknown. Previous studies suggested that IE1 residues 1 to 222 contribute to phosphorylation (44). Thus, we tested the role of the N-terminal replication domain in IE1 phosphorylation. When produced in uninfected, plasmid-transfected SF21 cells, wild-type IE1 migrated as a doublet on high-resolution SDS-polyacrylamide gradient gels (Fig. 5A, lane 2). Treatment with calf-intestinal phosphatase (CIP) reduced or eliminated the slower-migrating species (lane 3), whereas phosphatase inhibitor EDTA did not (lane 4). This finding confirmed the CIP-sensitive phosphorylation of IE1 (3, 44). IE1(Δ24–32)^HA, which lacks residues flanking the replication domain, also exhibited an electrophoretic doublet that was CIP sensitive (lanes 14 to 16). In contrast, the replication-defective deletions IE1(Δ2–7)^HA, IE1(Δ8–15)^HA, and IE1(Δ16–23)^HA migrated as a single species; neither CIP nor EDTA affected their electrophoretic mobilities (Fig. 5A, lanes 5 to 13). This finding suggested that these IE1 deletions were hypophosphorylated and that residues 2 to 23 are required for IE1 phosphorylation. Alternatively, removal of residues 2 to 23 may have induced hyperphosphorylation of IE1 that was insensitive to CIP. The latter possibility is unlikely, because residues 2 to 32 are sufficient for phosphorylation of heterologous proteins (see below).

Fig 5.

IE1 phosphorylation. (A) Phosphatase treatment. Equal aliquots of soluble extract from SF21 cells transfected 24 h earlier with plasmids carrying genes for wt IE1^HA or the indicated deletions were treated with calf intestinal phosphatase (+), 50 mM EDTA (E), or 1% SDS–1% βME (−) and subjected to anti-HA immunoblot analysis by using high-resolution polyacrylamide gradient gels. A contiguous immunoblot (lanes 1 to 16) is shown. The larger, hyperphosphorylated species of IE1^HA is marked (P) in panels A, B, and D. (B) Phosphatase inhibition. SF21 cells were transfected with plasmid carrying the gene for wt IE1^HA or vector alone (−) and treated 24 h later with DMSO vehicle containing the indicated concentrations of okadaic acid in growth medium. Total cell lysates were prepared 2 h after drug treatment, separated by high-resolution SDS-PAGE, and subjected to anti-HA immunoblot analysis. (C) Schematic of IE1-GFP^HA fusions. N-terminal IE1 residues (open box) were fused to GFP with a C-terminal HA epitope tag (black box). (D) Phosphatase treatment of IE1-GFP^HA fusions. Soluble extracts obtained from SF21 cells transfected 24 h earlier with plasmid carrying the GFP^HA or the GFP^HA fusions containing the indicated IE1 residues were treated with phosphatase (+), EDTA (E), or SDS-βME (−) as described for panel A and subjected to anti-HA immunoblot analysis. Size standards (in kilodaltons) are indicated.

To investigate the dynamics of IE1 phosphorylation, we tested the in vivo effects of okadaic acid. This inhibitor of serine/threonine phosphatases blocks protein dephosphorylation and simultaneously stimulates intracellular kinases (5). Upon treatment of plasmid-transfected SF21 cells with increasing doses of okadaic acid, the slower-mobility hyperphosphorylated species of IE1 increased and the faster-mobility species declined (Fig. 5B, lanes 2 to 7). Only the hyperphosphorylated species was detected at inhibitor concentrations of 50 nM or higher (Fig. 5B, lanes 6 and 7). Immunoprecipitated IE1 from plasmid-transfected cells was also detected by phospho-specific staining (Pro-Q Diamond [45]) (data not shown), thus verifying IE1 phosphorylation. We concluded that the state of IE1 phosphorylation varies, depending on intracellular conditions (see below).

To determine if the IE1 replication domain is sufficient to confer phosphorylation, we constructed expression plasmids carrying genes for N-terminal IE1 residues that were fused to HA-epitope tagged GFP (Fig. 5C). Because GFP itself is phosphorylated poorly (see below), this strategy allowed us to assess alterations in the phosphorylation state of the fusion protein. When produced in plasmid-transfected SF21 cells, GFP^HA alone resolved as a single species that was unaffected by CIP or EDTA (Fig. 5D, lanes 2 to 4). In contrast, fusion of GFP to the N-terminal 32 or 64 residues of IE1 promoted phosphorylation. The larger, less-abundant species of IE1(1–32)GFP^HA and IE1(1–64)GFP^HA were sensitive to CIP but not EDTA treatment (Fig. 5C, lanes 11 to 16). Whereas IE1(1–15)GFP^HA displayed low-level phosphorylation (lanes 8 to 10), IE1(1–7)GFP^HA did not (lanes 5 to 7). We concluded that residues that included the IE1 replication domain were sufficient to confer phosphorylation upon a heterologous protein and that these same residues may contain a phosphorylation site(s).

Potential phosphorylation sites in the IE1 replication domain are required for virus multiplication.

Alignment of the N termini of IE1 homologs encoded by lepidopteran-specific alphabaculoviruses revealed a remarkable degree of conservation among residues in the replication domain that represent potential phosphorylation sites (Fig. 6A). In particular, we discovered that the sequence TPXR/H (where X is any residue), which is located approximately 10 to 20 residues from the IE1 N terminus, is invariant. Moreover, this sequence is similar to the consensus phosphorylation site (T/SPXR/K) for cyclin-dependent kinases of insects and mammals (reviewed in reference 34). Thus, to investigate the role of this motif and surrounding residues on IE1 function, we tested amino acid substitutions of potential phosphorylation sites (Fig. 6B). These mutations were evaluated first for complementation of the ie-1 KO bacmid. Substitutions IE1^T11A/S12A and IE1^S14A/T15A failed to complement virus production (Fig. 7A). In contrast, IE1^T2A, IE1^S9A/Y10F, IE1^S17A/S20A, and IE1^SYS24AFA were comparable to that of wild-type IE1. Immunoblot analysis confirmed that each IE1 substitution was produced at comparable levels (Fig. 7B). The replication-competent IE1 substitutions also stimulated P35 synthesis at levels comparable to that of wild-type IE1^HA (Fig. 7B, lanes 2 to 4, 8, and 9). Finally, we substituted an alanine for Thr15 within the TPXR/H motif (Fig. 6). Remarkably, IE1^T15A failed to support virus multiplication, despite normal protein accumulation (Fig. 7). Thus, a single amino acid substitution within the putative kinase motif was sufficient to abolish IE1 replicative function.

Fig 6.

Functional residues comprising the IE1 replication domain. (A) Alignment of IE1 N termini. AcMNPV IE1 residues 1 to 32 (top) were aligned with corresponding IE1 residues from the indicated baculoviruses. Similar residues are shaded. The invariant TPXR motif (AcMNPV IE1 residues 15 to 18) is outlined. X, any residue; ψ, hydrophobic residue. Virus abbreviations are those listed by the International Committee of Taxonomy of Viruses (www.ictvdb.org). (B) Function of N-terminal IE1 substitutions. (Top) The 32-residue replication domain (shaded) is shown relative to the transactivation domain (open) of AcMNPV IE1. (Bottom) Double or single alanine or phenylalanine substitutions are indicated (arrows) within the AcMNPV IE1 replication domain; substituted wild-type residues are overlined. Loss-of-function substitutions (asterisks) and substitutions without obvious changes to IE1 function (diamonds) are indicated (see the text).

Fig 7.

Potential IE1 phosphorylation sites required for AcMNPV multiplication. (A) Budded virus yields. SF21 cells were transfected with ie-1 knockout bacmid AcBacIE1KO (1 μg) alone or with plasmid (5 μg) carrying wt IE1^HA or the indicated IE1^HA substitutions (Fig. 6B) or no IE1 plasmid (none). Extracellular budded virus was quantified 72 h after transfection by TCID₅₀ measurements as described in the legend to Fig. 2B. Reported values are the averages ± standard deviations of virus yields (in PFU per ml) from three independent transfections. (B) Immunoblots. Whole-cell lysates prepared 72 h after transfection of SF21 cells as described for panel A were subjected to immunoblot analysis by using anti-HA (top), anti-P35 (middle), or anti-actin (bottom) antibody. Protein size standards (in kilodaltons) are indicated.

Potential phosphorylation sites are required for IE1-mediated DNA replication.

To explore the connection between IE1 phosphorylation and DNA replication, we tested the capacity of the IE1 substitutions to support hr origin-specific plasmid replication. Quantitation of replicated plasmid levels in transfected cells revealed that each of the complementation-deficient substitutions IE1^T11A/S12A, IE1^S14A/T15A, and IE1^T15A failed to support DNA replication (Fig. 8A). In contrast, the complementation-competent substitutions supported DNA replication as well as wild-type IE1. The protein level of each IE1 substitution was comparable to that of wild-type IE1 (data not shown). Thus, residues Thr11, Ser12, or Ser 14 and Thr15 within the TPXR/H motif are required for IE1's DNA replication activity. Of these residues, Thr15 is absolutely conserved among all alphabaculoviruses, whereas residues Thr11, Ser12, and Ser 14 are not (Fig. 6). Nonetheless, these less-conserved but essential residues likely contribute indirectly to phosphorylation within the TPXR/H motif. None of these residues were required for transcriptional activation by IE1, as transactivation by all of the IE1 substitutions was comparable to that of wild-type IE1 (Fig. 9B). We concluded that the replicative and transactivation functions of IE1 are separable.

Fig 8.

DNA replication and transcriptional activation by IE1 N-terminal substitutions. (A) DNA replication. SF21 cells were cotransfected with hr5-containing template plasmid, plasmids for constitutive expression of AcMNPV replication factors, and a plasmid carrying wt IE1^HA or the indicated IE1^HA substitution or no IE1 plasmid, as described in the legend to Fig. 4. Replicated plasmid DNA from duplicate transfections along with linearized hr5 template plasmid (template) was visualized by Southern blot analysis (top). Values reported (bottom) are the averages ± standard deviations of replicated DNA obtained from duplicate transfections, normalized to that of cells transfected with wild-type IE1^HA-carrying plasmid. (B) Transactivation assays. Extracts were prepared 24 h after transfection of SF21 cells with the indicated reporter plasmid (1 μg) alone (none) or with plasmid (1 μg) carrying wt IE1^HA or the indicated IE1^HA substitutions and assayed for active luciferase as described in the legend to Fig. 3C. The values reported are the average luciferase activities ± standard deviations from triplicate transfections.

Fig 9.

IE1 phosphorylation during AcMNPV infection of permissive and nonpermissive cells. Monolayers of Spodoptera SF21 (A) and Drosophila DL-1 (B) cells were mock infected (mi) or inoculated with AcMNPV recombinant (+) wt/lacZ (MOI, 10). The overlay was replaced 3 h later with supplemented medium containing either DMSO vehicle alone (−) or 10 mM caffeine (+). The cells were harvested at the indicated times (in hours) after inoculation, boiled immediately in 1% SDS–1% βME, electrophoresed on a 4-to-12% polyacrylamide gradient gel, and subjected to immunoblot analysis by using anti-IE1 serum, which detects both IE1 and the related IE0. Arrows denote the hyperphosphorylated (P) and hypophosphorylated forms of IE1.

IE1 phosphorylation is temporally regulated during infection.

We hypothesized that phosphorylation of the replication domain modulates IE1 function. If so, the state of IE1 phosphorylation should vary depending on the principal in vivo activity of IE1. Thus, we monitored phosphorylation during the course of AcMNPV infection. Using IE1 antiserum, both IE1 and the closely related IE0 were detected at early and late times after inoculation of SF21 cells (Fig. 9A). The faster-mobility (hypophosphorylated) IE1 species predominated from 0 to 6 h (lanes 2 and 3). However, hyperphosphorylated IE1 increased dramatically between 6 and 12 h and was maintained through 24 h (lanes 5 and 7). Although less abundant, IE0 exhibited a similar temporal pattern of phosphorylation; IE0 also contains the complete replication domain near its N terminus. Confirming IE0 phosphorylation, the slower-mobility IE0 species was sensitive to treatment with CIP (data not shown). Thus, the rate of increase in phosphorylated IE1 and IE0 was greatest during the period (6 to 12 h) in which virus DNA replication initiates and increases exponentially (42).

This pattern of hyperphosphorylation that coincided with the onset of DNA replication was consistent with activation of a host- or virus-encoded kinase(s). Alternatively, IE1 (and IE0) dephosphorylation might be inhibited during this period. To distinguish these possibilities, we tested the effect of caffeine on IE1/IE0 phosphorylation. Caffeine is a broad but effective inhibitor of kinases involved in cell cycle checkpoints (reviewed in reference 48). When added at the beginning of infection, 10 mM caffeine markedly reduced the level of hyperphosphorylated IE1 by 12 h, but it had little if any effect on the level of hypophosphorylated IE1 (Fig. 9A, compare lanes 5 and 6). This finding suggested that a caffeine-sensitive kinase is directly or indirectly responsible for IE1 phosphorylation.

To determine whether IE1 is subject to phosphorylation in an unrelated insect species, we used cultured Drosophila DL-1 cells. Upon AcMNPV inoculation, IE1 but not IE0 was produced in these nonpermissive cells (Fig. 9B). Although the level of viral DNA synthesis is lower, its timing is comparable to that of permissive cells (23, 43). Whereas hyperphosphorylated IE1 was the predominant species in DL-1 cells (Fig. 9B), its level increased dramatically (lane 5) with the onset of AcMNPV DNA replication between 6 and 12 h (43). Moreover, caffeine preferentially reduced the level of hyperphosphorylated IE1 (Fig. 9B). We concluded that a host or viral kinase that is most active during virus DNA replication mediates IE1 phosphorylation in dipteran and lepidopteran cells. This finding is consistent with regulation of IE1 replicative activities in a time-dependent manner by phosphorylation.

DISCUSSION

Large DNA viruses, including the baculoviruses, encode multifunctional transcription factors that are essential for virus multiplication. The molecular mechanisms by which these regulators carry out their diverse activities are of significant interest. IE1 is a potent activator of baculovirus transcription that also participates directly in DNA replication and is thus required for virus propagation. Here we report that IE1 encoded by AcMNPV possesses an N-terminal domain that is required for virus genome replication but is dispensable for transcriptional activation. This 23-residue replication domain contributed positively to IE1 phosphorylation in a temporally regulated manner during infection. Importantly, a conserved TPXR/H cyclin-dependent kinase-like motif within the replication domain was essential for IE1's DNA replication functions. Thus, our study suggests that the diverse multiplicative functions of IE1 are regulated by a mechanism involving phosphorylation that is conserved among the alphabaculoviruses.

Required functions encoded by an N-terminal replication domain of IE1.

By using a sensitive bacmid rescue assay, we determined that the N-terminal 23 residues of AcMNPV IE1 are necessary for virus multiplication (Fig. 2). These residues were shown to have no direct role in transcriptional activation by IE1 as judged by hr-dependent and hr-independent reporters (Fig. 3). Nor did these residues contribute to hr-specific DNA binding by IE1, as indicated by hr-specific binding assays (data not shown) (38). Rather, transient origin-specific DNA replication assays using a minimal complement of AcMNPV-encoded replication factors demonstrated that the N-terminal residues are necessary for IE1-mediated DNA replication (Fig. 4 and 8) and thus constitute a replication domain, as defined here. Our findings are therefore consistent with a similar role for OpMNPV IE1's N-terminal residues, which are required for origin-specific plasmid replication in transfection assays (32). Interestingly, the N terminus of AcMNPV IE1 and OpMNPV IE1 share few common residues (Fig. 6A), except for those in the putative phosphorylation motif (see below).

Contribution of the replication domain to IE1 phosphorylation.

Extending previous studies (3, 44), we report here that IE1 is phosphorylated in AcMNPV-infected cells of lepidopteran or dipteran origin (Fig. 5 and 9). Previous mapping suggested that IE1 residues 1 to 168 include those necessary for phosphorylation (44). Here, we have shown that residues comprising the replication domain of IE1 contribute to phosphorylation. Deletions of the replication domain (residues 2 to 23) caused loss of CIP-sensitive phosphorylation of IE1 (Fig. 5A). When fused to a heterologous protein, these same IE1 residues were sufficient to confer phosphorylation upon the fusion protein (Fig. 5C). These findings indicated that the replication domain is required for IE1 phosphorylation and may serve as the site of phosphorylation. Importantly, substitution of four potential phosphorylation sites (Thr11, Ser12, Ser14, and Thr15) within the replication domain disrupted the replicative functions of IE1 as judged by the results of bacmid rescue and origin-specific DNA replication assays (Fig. 7A and 8A). The same residues made no contribution to transcriptional activation by IE1 (Fig. 8B). These findings suggest that IE1 phosphorylation of the replication domain is required for IE1's DNA replication activity. Furthermore, they confirm that the replicative and transcriptional activation functions of AcMNPV IE1 are genetically separable, like the activators of other DNA viruses (see below).

Essential residues comprising the replication domain of AcMNPV IE1 (Fig. 6) lie within or adjacent to sites typically recognized by cdc2/cyclin B kinases (T/SPXR) and mitogen-activated protein kinases (PXTP) (49). These kinases play essential roles in cellular DNA replication, control of the cell cycle, the DNA damage response, and other crucial biological processes (reviewed in references 34 and 51). Significantly, we identified a TPXR/H kinase-like motif, positioned 10 to 20 residues from the N terminus (Fig. 6A), in 34 of 34 IE1 proteins of alphabaculoviruses for which sequence information was available. The TPX residues were invariant, whereas the fourth residue was an R (55%), H (21%), Q (15%), or K (6%), in decreasing frequency. The TPXR/H motif is embedded within the AcMNPV IE1 replication domain, which generally lacks sequence similarity with other IE1 N termini. This observation, when combined with the finding that residue Thr15 of the motif is essential for IE1's replicative activity (Fig. 7 and 8), argues that a host or viral kinase regulates IE1 function. Interestingly, cdc2 kinase activity increases in AcMNPV-infected Spodoptera cells starting 6 h after infection (1), which coincides with the dramatic increase in IE1 phosphorylation and the onset of viral genome replication. Further studies will be required to determine those IE1 residues that are phosphorylated, including those within the TPXR/H motif, and to identify the critical kinase(s) responsible for this regulatory modification (see below). As indicated by okadaic acid treatment (Fig. 5B), IE1 is phosphorylated in uninfected cells and likely exists in dynamic equilibrium between hyper- and hypophosphorylated forms. Thus, IE1 is susceptible to phosphorylation by cellular kinases. Nonetheless, the sudden increase in IE1 phosphorylation from 6 to 12 h after infection (Fig. 9) suggests the appearance of a virus-encoded kinase or the activation of a host kinase.

Mechanism by which phosphorylated IE1 regulates virus DNA replication.

We have shown that the replication domain, including Thr15 of the TPXR/H motif, is required for AcMNPV IE1's replicative activity and phosphorylation (Fig. 4 and 8). During infection, maximal phosphorylation of IE1 coincided with the onset and rapid acceleration of AcMNPV DNA replication (Fig. 9A). Caffeine, a broad-spectrum inhibitor of cdc/cyclin B kinase and DNA damage response kinases, dramatically reduced IE1 phosphorylation during this period. Importantly, caffeine blocked AcMNPV DNA replication (13; J. Mitchell and P. Friesen, unpublished data) and thus demonstrated the link between phosphorylation and IE1's replicative activity. These findings suggest that phosphorylation regulates IE1's participation in virus genome replication.

A variety of mechanisms can be used by DNA sequence-specific transcriptional activators to regulate virus DNA replication (reviewed in reference 24). Such mechanisms include (i) direct recruitment of viral and host replicative factors to the origins of DNA replication, (ii) site-specific alteration (melting) of the DNA structure to promote initiation of DNA synthesis, and (iii) chromatin remodeling by activation of promoters near origins of DNA replication. Our study here is consistent with a model in which IE1 functions by nucleating the assembly of the DNA replication complex via binding to viral hr origins and recruiting requisite replication factors. Hyperphosphorylation, which coincided with the initiation of virus DNA replication, may activate IE1's capacity to recruit replicative factors to the hr origins of DNA replication. The acquisition of negatively charged phosphates may facilitate direct factor binding to the IE1 replication domain or trigger a conformational change that acts similarly. Indeed, IE1 associates with several baculovirus replication factors, including single-stranded DNA-binding protein LEF-3 and DNA helicase P143, and it colocalizes with these factors within nuclear DNA replication structures (15, 25, 28). This type of mechanism, which depends on functional modification by phosphorylation, is attractive because it provides a means by which IE1 may switch from transcriptional activation to DNA replication by exchanging auxiliary transcription factors for necessary DNA replicative factors at the origin in a temporally controlled manner. Additional studies are necessary to investigate this interesting possibility.

Common mechanisms regulating DNA replication by viral transcription activators.

The unexpected presence of the TPXR/H motif at the N terminus of most if not all IE1 proteins suggests that the alphabaculoviruses use a highly conserved mechanism for IE1-mediated DNA replication. Furthermore, IE1 shares multiple characteristics with other viral transcriptional regulators, including Epstein-Barr virus (EBV) Zta, human papillomavirus (HPV) E2, and polyomavirus large T antigen. The N-terminal residues of Zta are necessary for virus DNA replication but are dispensable for transcriptional activation (reviewed in reference 16); these Zta residues are required for recruitment of viral helicase (BBLF4), primase (BBLF1), and primase-associated factor (BBLF2/3). The N terminus of HPV E2 also encodes a separable replication domain (40) that is required for interaction with the HPV helicase E1 (reviewed in reference 12). The N-terminal residues of simian virus 40 large T antigen contain a J (DnaJ chaperone) domain, which is required for virus DNA replication and transcriptional control (reviewed in reference 47). The J domain interacts with and activates heat shock protein Hsc70, which may promote the proper arrangement of replicative factors within the replication complex. Importantly, the replicative functions of T antigen are also regulated by phosphorylation (reviewed in reference 14). Our finding that the functional organization of baculovirus IE1 resembles that of transcriptional regulators of diverse vertebrate viruses suggests that the mechanistic strategy of embedding replication-specific functions at the N terminus is an effective and conserved feature of DNA viruses of mammals and insects.