Identification of Six Autographa californica Multicapsid Nucleopolyhedrovirus Early Genes That Mediate Nuclear Localization of G-Actin

Abstract

Nuclear filamentous actin (F-actin) is required for nucleopolyhedrovirus (NPV) progeny production in NPV-infected, cultured lepidopteran cells. We have determined that monomeric G-actin is localized within the nuclei of host cells during the early stage of infection by Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV). With a library of cloned AcMNPV genomic fragments, along with a plasmid engineered to express enhanced green fluorescent protein-Bombyx mori G-actin in transient transfection experiments, we identified six AcMNPV early genes that mediate nuclear localization of G-actin in TN-368 cells: ie-1, pe38, he65, Ac004, Ac102, and Ac152. Within this subset, ie-1 and pe38 encode immediate-early transcriptional transactivators, he65 encodes a delayed-early product, and the products encoded by Ac004, Ac102, and Ac152 have not been characterized. We found that when driven by foreign promoters, ie-1, pe38, and Ac004 had to be expressed prior to Ac102 or he65 for nuclear G-actin to accumulate and that expression of Ac152 was no longer required. These results and others suggested that the product of Ac152 was a transactivator (directly or indirectly) of both Ac102 and he65 and that recruitment of G-actin to the nucleus was a temporally regulated process. Determining the functions of each of the six AcMNPV gene products with respect to our assay should provide valuable clues to basic cellular mechanisms of actin regulation and how AcMNPV infection affects them.

Members of the Nucleopolyhedrovirus genus of the Baculoviridae are divided into two groups based on the phylogenetic relatedness of their double-stranded DNA genomes (16). Members of both groups, however, are dependent upon nuclear filamentous (F) actin for progeny production, suggesting that this unique trait is ancestral to the nucleopolyhedrovirus (NPV) lineage (19). The best-studied baculovirus is Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV), a widely used expression vector and the type species of the NPVs (4, 21, 39, 51). Not surprisingly, AcMNPV has led the way in studies of baculovirus manipulation of the actin cytoskeleton, and these studies have brought to light a number of features regarding this process.

The budded form of AcMNPV (AcMNPV BV) is taken into cells by adsorptive endocytosis, and nucleocapsids penetrate the cytoplasm when the viral envelope fuses with the endosomal membrane (53). Fusion is triggered by the decreasing pH of the endosome and is mediated by the BV envelope protein, gp64 (6). Immediately following penetration, F-actin cables can be detected within the cytoplasm, often in association with viral nucleocapsids (10). These actin cables are thought to facilitate transport of the nucleocapsids to the nucleus (29), where genome replication and nucleocapsid morphogenesis take place (12). Following uncoating within the nucleus and expression of early viral genes, the actin cytoskeleton assumes a second configuration, with the formation of F-actin aggregates at the plasma membrane (9, 48). This second change is mediated by the product of a single early viral gene, arif-1, which is a ∼47-kDa phosphoprotein that colocalizes with the F-actin aggregates (11).

During late gene expression, a third and more dramatic shift in the actin cytoskeleton takes place: F-actin appears within the nucleus, both within the central virogenic stroma, where viral DNA synthesis occurs (58), and also (and more prominently) within the surrounding “ring zone,” juxtaposed to the inner nuclear membrane (9, 52, 54). Within the nucleus, p78/83, a minor capsid protein, colocalizes with the F-actin (28), and p39, the major capsid protein, partially colocalizes with it (9). Both p78/83 and p39 are actin-binding proteins; p78/83 appears to bind F-actin preferentially, while p39 binds both F- and G-actin (29)

Nuclear F-actin is essential for nucleocapsid morphogenesis (38, 52). Accordingly, cytochalasin D and latrunculin A, F-actin-disrupting drugs, prevent progeny virus production (19). In the presence of cytochalasin D, viral DNA is not packaged properly, nor are capsids produced in unit lengths (52). Instead, long and predominantly empty tubules, replete with p39, accumulate, and they are not associated with basal structures. In contrast, normal AcMNPV nucleocapsid morphogenesis begins with the formation of cylindrical capsid sheaths (presumably composed of the major capsid protein, p39) in association with basal structures (12). Basal structures of Orgyia pseudotsu-gata MNPV have been reported to contain the O. pseudotsu-gata MNPV homologue of p78/83 (49), and by inference, the basal structures of AcMNPV should contain p78/83. If so, then it is plausible that F-actin is involved in the assembly process during the earliest stages of nucleocapsid morphogenesis. The recent finding that p78/83 and its homologues contain putative Wiskott-Aldrich syndrome protein sequence motifs that bind and activate the Arp2/3 complex (a eukaryotic multifunctional nucleator and organizer of actin [25]) supports this speculation (35).

Whatever the mechanism, in order for actin to facilitate nucleocapsid morphogenesis within the nucleus, it first must be recruited there. Here we report that monomeric G-actin accumulated within nuclei of AcMNPV-infected cells during the early stages of infection and that expression of six AcMNPV early genes was sufficient for this to occur in transiently transfected TN-368 cells. Two of the six genes involved in the nuclear localization of G-actin, he65 and Ac102, were found to have overlapping functions, and the product of a third gene, Ac152, was needed to activate expression of he65 and Ac102. When driven by foreign promoters, expression of only four of the nuclear localization of G-actin genes, ie-1, pe38, Ac004, and either Ac102 or he65, resulted in the accumulation of G-actin within nuclei of TN-368 cells.

While the existence of actin, actin-related proteins, and actin-binding proteins within the nuclei of metazoan cells has historically been controversial, it is now widely accepted, but the roles of these proteins within nuclei are still unclear (43). Identification of the AcMNPV nuclear localization of G-actin genes described herein not only advances our understanding of baculovirus biology but also may help to elucidate the cellular processes involved in the transport, retention, regulation, and function of nuclear actin.

MATERIALS AND METHODS

Cells and viruses.

Sf21 cells (13) were cultured at 27°C in Grace's medium (JRH Biosciences) supplemented with 10% fetal bovine serum (Gibco-BRL). TN-368 cells (17) were cultured at 27°C in Grace's medium supplemented with 10% fetal bovine serum, lactalbumin (Gibco-BRL), and yeastolate hydrolysate (Gibco-BRL). Media for cells grown in suspension culture were supplemented with 0.1% pluronic F-68 (Sigma). The E2 strain of AcMNPV was used in this study (50).

Plasmids.

An AcMNPV E2 genomic plasmid library was assembled with the help of colleagues at Texas A&M University, Max Summers and Linda Guarino. EcoRI-A (in pUC19) was obtained from Max Summers, and EcoRI-B (in pUC8), HindIII-B (in pUC8), HindIII-F (in pUC8), PstI-C (in pUC8), PstI-F′-E′ (in pUC8), and hsp70 ie-1 (in pHSEH) were obtained from Linda Guarino. We cloned EcoRI-J into pBSKS+ (Stratagene).

Expression vector pIE1/153A (33) was supplied to us by Kostas Iatrou (University of Calgary). Bombyx mori cytoplasmic A4 actin DNA (38) was obtained from Nicole Mounier (University of Lyon), and pBShsp was provided by Gerald Rubin (University of California). eGFP, used for the eGFP-A4 gene fusion, originated from pEGFP-Actin (Clontech). pBSKS+ (Stratagene) and pUC18 (Pharmacia) were used for subcloning fragments of the AcMNPV genome.

Plasmid constructions.

Expression plasmids driving enhanced green fluorescent protein (eGFP)-B. mori A4 actin were constructed in several steps. First, the human actin sequence in the commercial vector pEGFP-Actin was replaced with the B. mori A4 actin gene. For in-frame fusion to eGFP, B. mori A4 actin was removed from a previously constructed pAcMP3-based plasmid (39) with XbaI and BglII, blunted, and inserted into the XhoI- and BamHI-blunted pEGFP-Actin vector. The orientation of the insert was determined by restriction analysis and DNA sequencing. The eGFP-B. mori A4 actin sequence was removed subsequently with NheI and XbaI and inserted into pIE1/153A cleaved with XbaI to form pIE1actGA.

The ie-1 gene was excised from the pIE1/153A-based plasmids that contained the NLA genes to create pact plasmids with NheI and ApaI followed by blunting and religating (Fig. 1). The pIE1/153A and the pact-based plasmids therefore differ functionally from one another in that the latter do not express IE-1 in addition to eGFP-actin (Fig. 1). Finally, PCR-amplified NLA genes were blunt ended and ligated into HindIII-cleaved, blunted pBShsp to create plasmids containing NLA genes driven by the Drosophila melanogaster hsp70 promoter. Orientation of the inserts in the plasmids was determined by restriction analysis and DNA sequencing. We found that when eGFP-B. mori actin was expressed in cultured insect cells, regardless of the vector, it polymerized and produced no apparent abnormal cellular effects, suggesting that the GFP-tagged actin was functional.

FIG. 1.

Vector constructs used in transfection experiments. Common plasmid vector backbone details are omitted. pIE1/153A and pIE1actGA contain the ie-1 open reading frame driven by the ie-1 promoter (ie-1 p). hr3, B. mori NPV homologous region 3; act p, B. mori A3 actin promoter; MCS, multiple cloning site; poly A, polyadenylation signal; eGFP-actin, fused open reading frames of enhanced green fluorescent protein and actin.

Subcloning of AcMNPV genomic plasmids.

HindIII-F was subcloned into smaller fragments PstI-N (2.7-kb PstI fragment of HindIII-F cloned into pBSKS+), SphI-J (5.4-kb SphI fragment of HindIII-F, cloned into pUC18), HindIII-FES (2.9-kb EcoRI/SalI fragment of HindIII-F, cloned into pBSKS+), and HindIII-FXH (2.4-kb XhoI/HindIII fragment of HindIII-F, cloned into pBSKS+). PstI-C was subcloned into smaller fragments PstI-CH1 (5.5-kb PstI/HindIII fragment of PstI-C, cloned into pBSKS+), KpnIE (2.0-kb KpnI fragment of PstI-C, cloned into pBSKS+), KpnI-D (5.8-kb KpnI fragment of PstI-C, cloned into pBSKS+), EcoRI-L (4.0-kb EcoRI fragment of PstI-C, cloned into pBSKS+), and PstI-CE4 (4.3-kb EcoRI fragment of PstI-C, cloned into pBSKS+).

PCR amplification of AcMNPV open reading frames.

Ac001, Ac002, Ac004, Ac101, Ac102, Ac149, Ac150, Ac152, and pe38 were PCR amplified from PstI-C or HindIII-F with 5′ primers with an 8-nucleotide sequence containing a BamHI site followed by 20 nucleotides of gene-specific sequence immediately preceding the start codon and 3′ primers with an 8-nucleotide sequence containing an XbaI site followed by 20 nucleotides of (complementary) gene-specific sequence which ended 50 nucleotides downstream of the stop codon. he65 and Ac107 were PCR amplified from PstI-C with 5′ primers with an 8-nucleotide sequence containing a BamHI site followed by 20 nucleotides of gene-specific sequence immediately preceding the start codon and 3′ primers with an 8-nucleotide sequence containing a BamHI site followed by 20 nucleotides of (complementary) gene-specific sequence which ended 50 nucleotides downstream of the stop codon. Amplified genes were then cloned into BamHI-cleaved pIE1/153A or BamHI- and XbaI-cleaved pIE1/153A and verified by DNA sequencing.

Transfections.

Cells (1.9 × 10⁵ Sf21 or 8 × 10⁴ TN-368) were plated onto 22-mm² coverslips (in six-well plates) in a volume of 250 μl and allowed to settle for 30 min. DNA for transfection was prepared with Endofree maxi-prep kits (Qiagen or Sigma). In infection experiments, medium was removed, and cells were infected at a multiplicity of infection of 10 for 1 h (with or without 5 μg of aphidicolin per ml), then transfected at 2 h postinfection. Transfections were performed with Cellfectin (Gibco-BRL) by the manufacturer's protocol, with 0.5 μg of pIE1actGA and 5 μg of viral fragments (in plasmids) per coverslip.

In experiments with aphidicolin, 5 μg/ml was added to transfection media. Coverslips were fixed 2 to 3 days posttransfection for 10 min in 2% paraformaldehyde in PHEM buffer (60 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)], 25 mM HEPES, 10 mM EGTA, 2 mM MgCl₂, pH 6.9), washed once in PHEM, stained with 4′,6′-diamidino-2-phenylindole (DAPI) (Sigma; 1 μg/ml in PHEM) for 10 min, washed twice more in PHEM, and mounted on microscope slides in Prolong antifade mounting medium (Molecular Probes). In some experiments, cells were solubilized following DAPI treatment with 0.15% Triton X-100 and stained with tetramethylrhodamine isothiocyanate-phalloidin (Sigma). Coverslips were viewed with a Zeiss Axiophot photomicroscope equipped for fluorescence microscopy or a Zeiss confocal laser scanning microscope.

RESULTS

Nuclear G-actin accumulation is mediated by early viral genes.

To determine whether the G-actin that polymerizes within the nuclei of AcMNPV-infected cells during late gene expression was recruited to the nucleus during early gene expression, we transfected AcMNPV-infected Sf21 cells with pIE1actGA at 2 h postinfection in the presence and absence of 5 μg of aphidicolin per ml, a drug that prevents AcMNPV late gene expression (45). Sf21 cells infected in the presence of aphidicolin showed eGFP-actin in the nucleus (Fig. 2A and B). In contrast to cells infected in the absence of aphidicolin, this nuclear eGFP-actin did not stain with rhodamine phalloidin, indicating that it was present in its unpolymerized, monomeric state (Fig. 2C). These results showed that early gene expression was sufficient for the nuclear localization of G-actin and confirmed the requirement of late gene expression for subsequent polymerization of nuclear G-actin (not shown) (9).

FIG. 2.

Fluorescence microscopy images of Sf21 cells infected with AcMNPV at a multiplicity of infection of 10 in the presence of 5 μg of aphidicolin per ml and then transfected with pIE1actGA at 2 h postinfection. Cells were fixed and stained at 48 h postinfection. (A) DAPI-stained DNA. (B) eGFP-B. mori G-actin. (C) Tetramethylrhodamine isocyanate-phalloidin staining of filamentous actin.

Transfection of pIE1actGA into TN-368 cells along with mixtures of viral fragments covering the AcMNPV genome.

The finding that AcMNPV early gene expression was sufficient to localize G-actin within the nucleus made it possible to carry out transient expression of AcMNPV genomic fragments along with pIE1actGA in order to identify early viral factors involved in this process. Baculovirus early genes are expressed exclusively by host factors alone (immediate-early genes) or by host factors in combination with immediate-early virus-encoded transactivators (delayed-early genes) (for a review, see reference 47).

We began with three mixtures of plasmids that contained between 24 and 36% of the AcMNPV genome (Fig. 3). These mixtures were transfected along with pIE1actGA into TN-368 cells, which were subsequently processed at 3 days posttransfection. Use of pIE1actGA eliminated the need to add a separate plasmid encoding IE1, a well-characterized transactivator of a number of AcMNPV delayed-early genes. We used TN-368 cells rather than Spodoptera frugiperda-derived cell lines for these experiments because TN-368 cells do not undergo IE1-induced apoptosis as readily as do cells in the S. frugiperda lineage (41).

FIG. 3.

Genomic locations of plasmid DNAs used in transfections with pIE1actGA. Three separate mixtures, represented on three separate lines, are shown beneath the EcoRI map of AcMNPV.

The mixtures containing either the HindIII-F, PstI-C, and HindIII-B fragments or only the HindIII-F and PstI-C fragments (covering approximately 20% of the genome) showed a pattern of eGFP-actin localization distinct from all other transfection mixtures (Fig. 4A and B). The majority of these transfected cells showed a high level of eGFP-actin signal that made it difficult to distinguish subcellular structures, but eGFP-G-actin was clearly localized within the nuclei of approximately 10% of the cells. In addition, cells showing obvious cytoplasmic but no nuclear signaling, such as those shown in Fig. 4C and D, were not found in this positive transfection mixture.

FIG. 4.

Fluorescence microscopy images of TN-368 cells transfected with cloned fragments of AcMNPV DNA, at 3 days posttransfection. (A) eGFP-B. mori G-actin in cells transfected with pIE1actGA and the PstI-C and HindIII-F fragments of AcMNPV. (B) DAPI staining of cells shown in panel A. (C) eGFP-B. mori G-actin in cells transfected with pIE1actGA and the EcoRI-A, EcoRI-J, and PstI-F′-E′ fragments of AcMNPV. (D) DAPI staining of cells shown in panel C.

Subcloning of the PstI-C and HindIII-F fragments to identify the viral genes involved in the nuclear localization of G-actin.

To identify specific viral genes involved in the nuclear localization of G-actin, we subcloned the PstI-C fragment into the PstI-CH1, KpnI-E, KpnI-D, EcoRI-L, and PstI-CE4 fragments (Fig. 5). We subsequently transfected each of these subclones in combination with HindIII-F and pIE1actGA into TN-368 cells and found that either the PstI-CH1, KpnI-E, or KpnI-D fragment gave a strong positive signal in combination with HindIII-F (on the order of 10% clearly positive cells and few to no clear negatives). KpnI-D and KpnI-E are distinct clones, and therefore we reasoned that these clones each contained at least one gene with apparent overlapping function that could complement HindIII-F in our assay.

FIG. 5.

Map showing subcloned fragments of PstI-C and HindIII-F. The locations of the ends of the fragments within the AcMNPV genome are indicated by the numbers flanking PstI-C and HindIII-F. Predicted open reading frames (2) are indicated below the PstI-C and HindIII-F fragments by number (and gene name if previously characterized), with the predicted direction of transcription indicated by arrowheads.

To specify which specific open reading frames (ORFs) were involved, we PCR amplified ORFs Ac101, Ac102, Ac103, Ac105 (he65), and Ac107 (Fig. 5) and inserted each into the pIE1/153A expression vector. Subsequent transfections of individual ORFs with HindIII-F and pIE1actGA showed that either Ac102 or Ac105 (he65) could act in combination with HindIII-F to induce the nuclear localization of G-actin.

Relative to PstI-C, the subcloning of the HindIII-F fragment proved to be more challenging. We subcloned HindIII-F into the PstI-N, SphI-J, HindIII-FES, and HindIII-FXH fragments, but no combination (even combinations with apparent complete coverage of HindIII-F) resulted in a clear positive signal when transfected with PstI-C. One possible explanation for this result was that a homologous region present within HindIII-F was lacking in some of the subclones containing ORFs involved in the nuclear localization of G-actin. The lack of a homologous region may have resulted in lower relative expression of those ORFs because homologous regions bind IE-1 and act as transcriptional enhancers (8, 15, 30, 37, 46).

Candidate ORFs, therefore, were PCR amplified from HindIII-F and cloned into pIE1/153A. AcMNPV ORFs Ac149, Ac150, Ac151 (ie-2), Ac152, Ac153 (pe-38), Ac002, and Ac004 were amplified and tested in combination with PstI-C. A clear positive result was obtained when the PstI-N fragment was transfected along with pe-38, Ac004, and PstI-C or when PstI-N, HindIII-FES, and pe-38 were transfected along with PstI-C. Further transfections showed that ORFs Ac152 and Ac004 in combination with pe-38 and PstI-C were sufficient for nuclear localization of G-actin (Fig. 6B).

FIG. 6.

Confocal fluorescence microscopy images of TN-368 cells transfected with pIE1actGA alone or with cloned genes or genomic fragments of AcMNPV. Cells were fixed 3 days posttransfection and stained with DAPI. (A) PstI-C and HindIII-F fragments of AcMNPV. (B) Ac004, Ac152, and pe-38 and 24 h later PstI-C. (C) HindIII-F and 24 h later Ac102. (D) HindIII-F and 24 h later he65. (E) Ac004, Ac152, and pe-38 and 24 h later Ac102. (F and G) Ac004, Ac152, and pe-38 and 24 h later he65. (H) pIE1act GA alone.

Timing of expression of genes involved in nuclear localization of G-actin.

Once Ac102 and he65 from PstI-C and Ac152, Ac004, and pe-38 from HindIII-F were identified as nuclear localization of G-actin (NLA) genes, we cloned each of them into pIE1/153A and subsequently transfected these plasmids into TN-368 cells. We were disappointed to find that this transfection mixture did not give a positive signal. Previous experiments in which PstI-C or HindIII-F was transfected with individual ORFs cloned into pIE1/153A had suggested that timing of expression of NLA genes was important. For example, we had noted that PstI-C along with Ac152, pe-38, and Ac004 gave a positive signal at 3 but not 2 days posttransfection unless ie-2, an immediate-early transactivator gene, was included in the mix.

We carried out a series of experiments, therefore, in which we transfected subsets of various combinations of the NLA genes 1 day in advance of the remainder and then processed the cells 1 or 2 days following the second transfection. We found that when Ac152, pe-38, and Ac004 were transfected 1 day prior to either Ac102 or he65 and the cells were processed 2 days after the second transfection, nuclear G-actin was clearly detected (Fig. 6E and F).

Reduced number of NLA genes required when driven by a foreign promoter.

The five NLA genes cloned into pIE1/153A yielded a clear positive signal when the appropriate timing was observed. Having established this, we conducted experiments to determine whether all five of the gene products were needed when NLA gene expression was driven by the cellular actin promoter in the pIE1/153A vector (31). We noted that mixtures containing pIE1actGA and PstI-C along with pe-38 and Ac004 (but not Ac152) cloned into pIE1/153A did not give a positive signal, whereas mixtures of pIE1actGA and pe-38, Ac004, and either Ac102 or he65 cloned into pIE1/153A did give a positive signal. These results indicated that the Ac152 gene product served to activate and/or enhance expression of both Ac102 and he65 when these genes were driven by their native viral promoters. Hence, expression of a minimum subset of three NLA genes cloned into pIE1/153A, pe38, and Ac004 plus Ac102 or he65 was sufficient for mediating G-actin accumulation within the nucleus (Fig. 5).

In order to determine whether IE1 was necessary when expression of the NLA genes was driven by the actin promoter in pIE1/153A, we deleted ie-1 from the pIE1/153A vector (creating pact [Fig. 1]), inserted each of the NLA genes and eGFP-B. mori A4 actin into the modified vector, and transfected TN-368 cells. We found that in the absence of IE1, we did not get a positive nuclear signal, but we were able to rescue the signal when we added a single vector that expressed ie-1 to the mix (e.g., pIE1actGA). These results suggested that the nuclear localization of G-actin activity depended on either enhanced expression of one or more of the NLA genes through the interaction of IE1 with the homologous region 3 (hr3) sequence in pIE1/153A (Fig. 1) (31), or on an interaction of IE1 with a cellular factor(s), or both.

To distinguish among these possibilities, we next cloned the NLA genes behind the Drosophila hsp70 promoter in the plasmid pBShsp. We transfected TN-368 cells sequentially with pactGA on day 1, pBShsp/ie-1, pBShsp/pe38, and pBShsp/Ac004 on day 2, and pBShsp/Ac102 on day 3; when we examined those cells on day 5, we found that G-actin had been recruited to the nucleus. In control experiments in which pBShsp/ie-1 was not added, no nuclear localization of G-actin was evident, indicating that IE1 was needed for interaction with cellular factors in our assay.

DISCUSSION

Actin is an abundant protein in all eukaryotes and is the major component of the microfilament network of the cytoskeleton. This protein has been studied for decades and has many well-defined roles, all of which are executed within the cytoplasm. Many viruses use actin in their transport, assembly and replicative processes, but with the exceptions of AcMNPV nucleocapsid assembly (38) and the export of human immunodeficiency virus type 1 RNAs (18, 20), these activities all occur within the cytoplasm as well. Our discovery that nuclear F-actin was required for nucleocapsid morphogenesis therefore suggested that AcMNPV gene products were involved in the recruitment and/or retention of actin within the nucleus. Here we report that localization of G-actin within the nucleus occurs during early viral gene expression and that the products of six early viral genes, ie-1, pe38, he65, Ac004, Ac102, and Ac152, are involved in this activity.

Among the products of these genes, IE1 is the best characterized. It is highly conserved among baculoviruses and is a transcriptional transactivator of several early, late, and very late genes (1, 5, 15, 30, 40), including an NLA gene, he65 (26). Our finding that IE1 was essential for nuclear localization of G-actin, even when expression of the other required NLA genes was independent of its presence, suggested that IE1's role in our assay was mediated through interactions with host factors. It is known that IE1 has the capability of interacting with host factors because its expression can induce apoptosis in Sf21 cells (41). Interestingly, IE1's ability to induce apoptosis is augmented by another NLA gene product, PE38 (42). It is possible, therefore, that initiation of the apoptotic process is part of the viral strategy for recruiting actin to the nucleus. In support of this idea, actin is recruited to nuclei of HL60 human promyelocytic cells induced to undergo apoptosis (34). In addition, during apoptosis, chromatin undergoes characteristic morphological changes involving margination toward the nuclear periphery, similar to chromatin changes that occur in baculovirus-infected cells (1, 56). Moreover, it has been postulated that in apoptosing cells, nuclear actin may play a role in chromatin margination (34).

IE1 and PE38 also have been implicated in viral DNA synthesis. In plasmid-based assays, IE1 is essential for DNA synthesis, and PE38 is stimulatory (22). In addition, IE1 binds to homologous regions which serve as origins of replication in plasmid-based assays (31). In infected cells, however, it is not known whether IE1 is needed for DNA replication beyond its role as an activator of required early genes, or if homologous region sequences act as origins of replication. Even so, existing data are consistent with the hypothesis that IE1 has a direct role in origin binding and replication complex formation (31). If so, then PE38 may also participate in this activity.

Not only does PE38 stimulate DNA synthesis in plasmid-based assays, but AcMNPV pe38 deletion mutants have been found to be defective in viral DNA synthesis and BV progeny production (D. Theilmann, personal communication; M. L. Milks, J. O. Washburn, L. G. Willis, L. E. Volkman, and D. Theilmann, submitted for publication). In addition, PE38 has two putative DNA-binding motifs, a RING motif, a leucine zipper, and a nuclear translocation signal (23, 24). A consistent theme among proteins with RINGs is their ability to mediate protein associations that lead to the formation of large multiprotein complexes (7), such as chromatin remodeling complexes (44). It is possible that IE1, PE38, and F-actin participate in such a complex. Consistent with this idea, parental and progeny AcMNPV DNAs assume chromatin-like structures early during infection (57), and nuclear actin and actin-related proteins (arps) have been identified as components of chromatin remodeling complexes (44).

We determined that, when transfected into TN-368 cells, a mixture of Ac004, pe38, and PstI-C was insufficient for induction of nuclear localization of G-actin. Transfection with ORFs Ac152, Ac004, and pe38 together with PstI-C, however, or pe38, Ac004, and either Ac102 or he65 cloned into pIE1/153A was sufficient for nuclear localization of G-actin (Fig. 6B). These results provided evidence that the product of Ac152 transactivated both Ac102 and he65. If this result is confirmed, then the products of Ac152 and ie-1 are both transactivators of the early gene he65 (3). Why both would be needed in the context of our assay is not apparent, but another AcMNPV delayed-early gene, p143, has also been found to be positively regulated by two different AcMNPV-encoded transactivators, IE1 and PE38 (30).

The finding that the products of Ac102 and he65 can function individually in combination with the other NLA gene products to recruit G-actin to the nucleus indicates that the products of these genes have overlapping functions, at least in TN-368 cells. The product of he65 is predicted to have zinc finger and DNA ligase motifs (not shown), but these motifs were not found in the Ac102 sequence, suggesting that the two gene products may function through different pathways or by different modes of action. AcMNPV has an extremely wide host range (14), and it is possible that these products have various degrees of activity in different hosts and/or host tissues. Species-specific effects and tissue-specific functions of AcMNPV gene products have been reported previously (5, 27, 32, 41).

Once the NLA genes were cloned into expression vectors, it became evident that timing of their expression was important for a clear nuclear signal to appear. We found that ie-1, pe38, and Ac004 had to be expressed prior to he65 or Ac102. The reasons for this are not clear at present, in part because the functions of the NLA gene products are largely unknown. Moreover, few clues to their function were gleaned through scanning the DNA and protein databases because none of the NLA sequences showed significant sequence identity beyond zinc finger and leucine zipper motifs. One activity that might be expected for members of the complex, however, is the masking of the two leucine-rich nuclear export signal sequences recently identified on actin (55). Whether or not the NLA gene products retain nuclear G-actin in this manner, the determination of their roles in our assay will lead to a better understanding of the cellular processes involved in the transport, retention, regulation, and function of nuclear actin.