|Home | About | Journals | Submit | Contact Us | Français|
Two types of viruses are produced during the baculovirus life cycle: budded virus (BV) and occlusion-derived virus (ODV). A particular baculovirus protein, FP25K, is involved in the switch from BV to ODV production. Previously, FP25K from the model alphabaculovirus Autographa californica multiple nucleopolyhedrovirus (AcMNPV) was shown to traffic ODV envelope proteins. However, FP25K localization and the domains involved are inconclusive. Here we used a quantitative approach to study FP25K subcellular localization during infection using an AcMNPV bacmid virus that produces a functional AcMNPV FP25K-green fluorescent protein (GFP) fusion protein. During cell infection, FP25K-GFP localized primarily to the cytoplasm, particularly amorphous structures, with a small fraction being localized in the nucleus. To investigate the sequences involved in FP25K localization, an alignment of baculovirus FP25K sequences revealed that the N-terminal putative coiled-coil domain is present in all alphabaculoviruses but absent in betabaculoviruses. Structural prediction indicated a strong relatedness of AcMNPV FP25K to long interspersed element 1 (LINE-1) open reading frame 1 protein (ORF1p), which contains an N-terminal coiled-coil domain responsible for cytoplasmic retention. Point mutations and deletions of this domain lead to a change in AcMNPV FP25K localization from cytoplasmic to nuclear. The coiled-coil and C-terminal deletion viruses increased BV production. Furthermore, a betabaculovirus FP25K protein lacking this N-terminal coiled-coil domain localized predominantly to the nucleus and exhibited increased BV production. These data suggest that the acquisition of this N-terminal coiled-coil domain in FP25K is important for the evolution of alphabaculoviruses. Moreover, with the divergence of preocclusion nuclear membrane breakdown in betabaculoviruses and membrane integrity in alphabaculoviruses, this domain represents an alphabaculovirus adaptation for nuclear trafficking of occlusion-associated proteins.
IMPORTANCE Baculovirus infection produces two forms of viruses: BV and ODV. Manufacturing of ODV involves trafficking of envelope proteins to the inner nuclear membrane, mediated partly through the FP25K protein. Since FP25K is present in alpha-, beta-, and gammabaculoviruses, it is uncertain if this trafficking function is conserved. In this study, we looked at alpha- and betabaculovirus FP25K trafficking by its localization. Alphabaculovirus FP25K localized primarily to the cytoplasm, whereas betabaculovirus FP25K localized to the nucleus. We found that an N-terminal coiled-coil domain present in all alphabaculovirus FP25K proteins, but absent in betabaculovirus FP25K, was critical for alphabaculovirus FP25K cytoplasmic localization. We believe that this represents an evolutionary process that partly led to the gain of function of this N-terminal coiled-coil domain in alphabaculovirus FP25K to aid in nuclear trafficking of occlusion-associated proteins. Due to betabaculovirus breakdown of the nuclear membrane before occlusion, this function is not needed, and the domain was lost or never acquired.
Insect-specific baculoviruses in the family Baculoviridae have a circular, double-stranded DNA genome of 88 to 180 kb with the capacity to code for 90 to 180 putative proteins (1). Based on the insect hosts from which these baculoviruses were isolated and their biological characteristics, they are divided into four genera: Alphabaculovirus, Betabaculovirus, Gammabaculovirus, and Deltabaculovirus (2). The alpha- and betabaculoviruses are composed of lepidopteran-specific nucleopolyhedroviruses (NPVs) and granuloviruses (GVs), whereas the gamma- and deltabaculoviruses are composed of hymenopteran- and dipteran-specific baculoviruses, respectively (2).
Baculoviruses, with the exception of gammabaculoviruses, undergo a biphasic virus life cycle during insect cell infection, initially involving the production of extracellular viruses or budded viruses (BVs) and later involving the production of intracellular viruses or occlusion-derived viruses (ODVs) (3, 4). After entry into the cell, early gene transcription by the host RNA polymerase II starts in the nucleus (3). Following viral DNA replication, late gene transcription by the viral RNA polymerase begins to produce viral structural proteins (5). Viral structural proteins and genomic DNA assemble to produce nucleocapsids in the virogenic stroma in the nucleus (6). Initially, nucleocapsids egress primarily through the nuclear pore complexes to the cytoplasm and acquire an envelope from the plasma membrane to bud out to produce BV (7, 8). In contrast, nucleocapsids destined to remain in the nucleus acquire an envelope from blebbing of the inner nuclear membrane (INM) to form ODVs (8, 9). This membrane-blebbing process is partially facilitated by a particular 25-kDa nucleocapsid protein, FP25K (8, 10, 11). During this late phase of infection, occlusion bodies (OBs) are formed around ODVs with either the matrix protein polyhedrin or granulin, in the case of alpha- and gammabaculoviruses or betabaculoviruses, respectively (3, 4, 8, 9). No polyhedrin/granulin orthologues are present in the deltabaculoviruses (4). Unlike other baculoviruses, the occlusion process of betabaculoviruses occurs after the breakdown of the nuclear envelope in a cytoplasmic-nuclear mix (3, 12, 13). The process of INM blebbing, facilitated by FP25K and other ODV envelope proteins, is thought to be a general switch from BV to ODV production (14,–16). This multifunctional FP25K protein is not required for infection in vitro, but it is necessary in vivo for proper ODV envelopment, virion occlusion, and insect per os infectivity (10, 16,–19).
Since it is not essential in vitro, and inactivation provides a selective advantage, the fp25k gene of baculoviruses readily undergoes mutation by transposition or replication slippage errors resulting in a base insertion or deletion (indel) during infection in cell culture (14, 20,–22). These mutations lead to an evident few-polyhedra (FP) phenotype, characterized by decreased production of OBs, fewer ODV envelope proteins localized to the nucleus, and increased BV titers (18, 22,–25). Therefore, it has been proposed that FP25K may be involved in associating with, sorting, and trafficking ODV envelope proteins to the INM and viral envelope, including ODV-E66 and ODV-E25 (17, 26, 27). Even though FP25K is not absolutely necessary for OB formation and ODV envelopment, its function yields more efficient localization of ODV envelope proteins to the nucleus, and it has been proposed that this nuclear trafficking is performed by other proteins, including E26 and importin-α-16 (26). In addition, there is a significant reduction in the major alphabaculovirus structural protein polyhedrin at both the transcriptional and protein levels in fp25k mutant infections (28). This suggests that FP25K may also play a role in the regulation of viral gene expression at the transcriptional level. Recent evidence also supports FP25K as a negative factor of BV infectivity due to increased GP64 abundance in fp25k mutant BV in addition to its ability to prohibit viral genome incorporation (29).
Most studies to date on FP25K have focused on its function in the context of alphabaculovirus infection rather than orthologues from other baculovirus genera. FP25K orthologues are present in alpha-, beta-, and gammabaculoviruses and absent in deltabaculoviruses (30). To gain insight into functional differences in alpha- and betabaculovirus FP25K proteins, Nakanishi et al. examined phenotypic changes between recombinant Bombyx mori NPV (BmNPV) expressing FP25K from Autographa californica multiple NPV (AcMNPV), Spodoptera litura multiple NPV (SpltMNPV), or Xestia c-nigrum GV (XecnGV) (19). They found that XecnGV FP25K was unable to rescue the defects in oral infectivity and postmortem host degradation observed in the BmNPV FP25K deletion mutant. It is possible that XecnGV FP25K is incompatible in this BmNPV; however, data from that study may also suggest that the fp25k functionalities of alpha- and betabaculoviruses are different. Even though fp25k orthologues were found in gammabaculoviruses, these fp25k sequences have low homology to alpha- and betabaculovirus fp25k sequences, with 8% identity and 14% similarity, and were therefore not considered for our localization studies (23, 31). The relatedness of this gammabaculovirus fp25k orthologue is still uncertain.
Previous studies on AcMNPV FP25K localization suggested that it is present near the endoplasmic reticulum (ER); however, to date, baculovirus FP25K localization and its amounts in the cytoplasm and nucleus have been inconclusive (26). Using immunogold transmission electron microscopy (TEM) analysis, Harrison and Summers revealed that AcMNPV FP25K localized to the cytoplasm but were unable to ascertain its presence in the nucleus due to the presence of electron-dense structures (32). Also, FP25K was confined to large “amorphous structures,” making its localization difficult to determine by traditional biochemical methods (32). In immunofluorescence studies, Braunagel et al. also found that FP25K was abundant in the cytoplasm, but its presence in the nucleus was not evident (17).
In order to study FP25K localization, we cloned an fp25k-gfp fusion gene into an fp25k insertion mutant bacmid virus [AcFP-GFP(ProPH)]. During AcFP-GFP(ProPH) infection, AcMNPV FP25K localized primarily to the cytoplasm, with a small percentage being localized in the nucleus. FP25K is predicted to contain a highly conserved coiled-coil domain (17). During infection with an N-terminal coiled-coil domain deletion virus (ΔL13:L44), FP25K localization significantly changed to be primarily nuclear. When Plutella xylostella granulovirus (PxGV) FP25K, which lacks the coiled-coil domain, was fused to green fluorescent protein (GFP) and cloned into the fp25k mutant bacmid, its infection also resulted in FP25K being localized predominantly to the nucleus. This suggests a critical role for this coiled-coil domain in maintaining alphabaculovirus FP25K cytoplasmic localization.
Trichoplusia ni BTI-TN-5B1-4 (Tn5), Spodoptera frugiperda IPLB-SF21AE (Sf21), and Sf21-specific clone (Sf9) cells were cultured in TNM-FH medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Atlanta Biologicals, Flowery Branch, GA) at 27°C. The AcMNPV bacmid (AcBacmid) viruses were produced by transfection of Tn5 cells using polyethylenimine (PEI) (Polysciences Inc., Warrington, PA) with bacmid DNA extracted from Escherichia coli strain DH10Bac (Invitrogen) (33). BV was harvested by collection of culture medium, followed by centrifugation at 500 × g for 5 min to remove cell debris. To generate a virus that produces nonfused GFP, the gfp coding sequence (CDS) was cloned from pBlueGFP into the pFastBac1 vector, and the resulting construct, pFastBac1-GFP, was transformed into competent DH10Bac bacteria to produce AcBacGFP in insect cells (10). Titers of all recombinant viruses, including viruses used for the setup of the time course study described below, were determined by endpoint dilution assays (50% tissue culture infective dose [TCID50]) as described previously (34). PxGV granules were provided by Q. Qin of the Chinese Academy of Science. Genomic DNA of PxGV was purified according to previously reported procedures (34). All primers referred to in this study are shown in Table 1.
Wild-type AcBacmid (bMON14272) DNA was isolated from the respective bacterial stocks and transfected into Sf21 cells. AcBacmid BV from transfection was harvested and subsequently passaged 10 times into fresh Sf21 cells by using the previously passaged BV as the inoculum. Viral DNA was extracted from passaged AcBacmid BV. PCR was performed on passaged AcBacmid BV DNA by using fp25k gene-specific primers fp25k-F and fp25k-R (Table 1) with expected 1.2-kb (wild-type) and 1.5-kb (mutant) fp25k product sizes (10). The BV of the passage that showed a prominent 1.5-kb PCR product (passage 10) was used for viral DNA extraction and eventual transformation of E. coli DH10B cells. Screening and selection of AcBacmid clones with an inactivated endogenous fp25k (AcBac-fp::287) were performed by colony PCR with the above-described fp25k PCR primers and conditions. The fp25k locus of AcBac-fp::287 was mutated through the insertion of a 287-bp piece of Sf21 host DNA into the fp25k coding sequence at the TTAA site at nucleotide (nt) +425, as previously reported (10). With a premature stop codon, this FP25K mutant sequence contains a 73-amino-acid (aa) deletion at the C terminus. fp25k::287 was sequenced, and the sequence is available under GenBank accession number KX284746. The resulting AcBac-fp::287 bacterial clone was transformed with helper plasmid pMON7124 to produce DH10Bac-fp::287 cells for ease of generation of recombinant viruses in our AcBac-fp::287 shuttle vector.
To generate AcFP-GFP(ProPH) and PxFP-GFP(ProPH), expressing GFP-tagged FP25K from AcMNPV and PxGV, respectively, the fp25k gene CDSs from AcMNPV and PxGV were PCR amplified by using the high-fidelity Pfu enzyme with genomic DNAs of the two viruses as the templates and cloned separately into pGEM-T Easy (Promega, Madison, WI) by using primers AcFP-fuse-F and AcFP-fuse-R for AcMNPV and primers PxFP-fuse-F and Px-fuse-R for PxGV (Table 1). To generate AcFP-GFP(ProFP), expressing GFP-tagged AcMNPV FP25K from its native fp25k promoter, the fp25k gene CDS and its fp25k promoter from AcMNPV were PCR amplified and cloned into pGEM-T Easy by using primers AcFPpromoter-F and AcFPpromoter-R.
For AcFP-GFP(ProPH) and PxFP-GFP(ProPH), the C-terminal ends of each of the fp25k CDSs were translationally fused to gfp in the pFastBac1 vector by BamHI/XbaI digestion of the fp25k cassettes and XbaI/XhoI digestion of the gfp cassette from pUC19-GFP (a clone collection in our laboratory), followed by subcloning into the BamHI/XhoI sites of pFastBac1, resulting in pFB1-AcFP-GFP(ProPH) and pFB1-PxFP-GFP(ProPH).
The fp25k promoter-driven bacmid [AcFP-GFP(ProFP)] was constructed differently. The fp25k CDS with the native fp25k promoter was translationally fused to gfp as detailed above; however, the BamHI/XhoI site of pFastBacHTa was used for subcloning, resulting in pFBHTa-AcFP-GFP(ProFP). In order to drive expression from only the native fp25k promoter, the polyhedrin (polh) promoter from this construct was deleted by SfoI/SnaBI digestion and self-ligation with T4 DNA ligase (New England BioLabs Inc., Ipswich, MA).
For the GFP-AcFP virus, gfp was translationally fused to the N terminus of AcMNPV fp25k. To do this, the gfp gene was PCR amplified from the pUC19-GFP template and cloned into pGEM-T Easy by using primers GFP-F-EcoRI and GFP-R-BamHI. The pFastBac1-GFP-AcFP vector was produced by EcoRI/BamHI digestion of the gfp cassette from pGEMT-GFP and BamHI/XbaI digestion of the fp25k cassette from pGEMT-Acfp25k (from a clone collection in our laboratory, which contains the AcMNPV fp25k open reading frame), followed by subcloning into the EcoRI/XbaI sites of pFastBac1.
All of these constructs were confirmed by sequencing and transformed into DH10Bac-fp::287 cells. The recombinant bacmid DNA produced was used to transfect Tn5 cells, and the resulting BV was harvested. Each of the resulting recombinant BVs was amplified only once in Tn5 cells to reduce any unwanted recombination of the fp25k-gfp fusion constructs.
AcMNPV fp25k was translationally fused to gfp as described above but subcloned into the pIE1 vector (an expression vector containing the ie-1 promoter). The AcMNPV fp25k fragment was recovered from pGEMT-Acfp25kΔTAA (produced as described above) by BamHI/XbaI digestion. The gfp fragment was cut out of pUC19-GFP by XhoI/Klenow digestion and then XbaI digestion. These fragments were subcloned into the XbaI/Klenow and BamHI sites of pIE1. The resulting construct, pIE1-AcFPGFP (at 1 μg), was transfected into Tn5 cells by using PEI and analyzed by confocal microscopy at 48 h posttransfection.
Tn5 cells (5 × 105) and Sf9 cells (1 × 106) in 35-mm tissue culture dishes were infected with AcBacGFP and AcFP-GFP(ProPH) at a multiplicity of infection (MOI) of 1 and harvested at 48 h postinfection (hpi). Cells were rinsed in phosphate-buffered saline (PBS), harvested in radioimmunoprecipitation assay (RIPA) buffer (25 mM Tris-HCl [pH 7.6], 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS]), and sonicated. The amount of protein in each sample was quantified by a Bradford assay (35). Each sample (30 μg) was heated at 100°C for 10 min in loading buffer (50 mM Tris-Cl [pH 6.8], 2% SDS, 0.1% bromophenol blue, 10% glycerol, and 100 mM dithiothreitol) and then subjected to SDS-polyacrylamide gel electrophoresis (PAGE). SDS-PAGE gels were either stained with Coomassie brilliant blue or transferred to nitrocellulose membranes (Schleicher & Schuell, Keene, NH) for Western blotting. The membrane was blocked with 5% nonfat dried milk (NFDM) in Tween plus Tris-buffered saline (TTBS) (150 mM NaCl, 50 mM Tris-HCl [pH 7.6], and 0.05% Tween 20). The membrane was treated with an anti-GFP antibody at a 1:2,000 dilution (Affinity Bioreagents, Golden, CO) overnight (16 h) at 4°C, washed three times with TTBS (without NFDM), and incubated with horseradish peroxidase (HRP)-linked anti-rabbit IgG at a 1:1,000 dilution (Cell Signaling, Danvers, MA) at room temperature (23°C) for 1 h. The membrane was washed three times with TTBS (without NFDM) and developed with HRP color development reagent (Bio-Rad, Hercules, CA). The same procedure was used for PxGV FP25K-GFP, although only Tn5 cells were infected with PxFP-GFP(ProPH) and AcBacGFP.
A time course was set up as follows: Tn5 cells (1 × 105) were infected in triplicate at an MOI of 1 with appropriate viruses, and BV was harvested at 12, 24, 36, 48, 60, and 72 hpi. The titers of all of the harvested BV samples were determined simultaneously by a modified quantitative real-time PCR (qPCR) method, as described previously (36). To convert the resulting viral copy number to PFU per milliliter, a standard virus, AcBacGFP, was used for qPCR. The titer of this standard virus was determined by a TCID50 assay. Instead of ie-1-specific primers for qPCR, we used primers pp34-F and pp34-R, which were resistant to primer-dimer formation (Table 1).
For all confocal analyses, either Tn5 cells (5 × 105) or Sf9 cells (1 × 106) were adhered to specialized 35-mm glass-bottom dishes (MatTek Co., Ashland, MA), and these cells were either left uninfected or infected with AcBacGFP or AcFP-GFP(ProPH) viruses at an MOI of 1. At 12, 24, 36, and 48 hpi, the cells were stained with a live, permeable nuclear stain (Hoechst 33342; Thermo Fisher, Waltham, MA) at a 1:1,000 dilution for 1 h at room temperature (23°C) and then washed twice and incubated with TNM-FH medium. Samples were imaged by using a confocal laser scanning microscope (FV500; Olympus) with a 40× or 100× oil immersion objective lens and FluoView software. To prevent bleed-through, sequential scans of the GFP and 4′,6-diamidino-2-phenylindole (DAPI) channels were used. Composite images were assembled by using Adobe Photoshop CS5 software. For z-stack imaging, 40 scans of both GFP and DAPI channels were taken at increments of a 500-nm depth.
To determine the percentages of GFP localized to the nucleus and cytoplasm, individual cells in the confocal images were manually analyzed by using ImagePro software. A total of 30 GFP-positive cells from each infection were chosen for quantitative measurements. Cells were selected based on their middle position within the focal plane. The middle position can be approximated by the near-round shape of the nuclei and condensation of DNA as seen in Hoechst 33342 dye-only images. By using ImagePro software, the total GFP sum intensity or the sum of all the pixels in a given region of each individual cell was determined in the GFP-only image. In the Hoechst 33342 dye-only image, a nuclear boundary was made for each cell based on the blue stain. This boundary was used in the GFP-only image to quantify the GFP sum intensity in the nucleus of that cell. The total GFP sum intensity of the cell was subtracted by the quantity in the nucleus to determine the amount of GFP in the cytoplasm. Therefore, the amount of GFP in either the nucleus or cytoplasm was divided by the total amount of GFP in the cell to determine the percentage in each cell compartment.
Tn5 cells (1 × 105) and Sf9 cells (5 × 105) were seeded into 24-well plates and infected with AcFP-GFP(ProPH) and AcFP-GFP(ProFP) at an MOI of 1 in triplicate. Infected cells were harvested at 48 hpi. Cells were rinsed once with PBS and lysed with 0.1% SDS at 27°C for 15 min. Lysis was confirmed by microscopy, and supernatants were transferred to 96-well black polystyrene plates (Greiner Bio-One, Monroe, NC). GFP measurements were taken at an excitation wavelength of 485 nm and an emission wavelength of 535 nm on a Filtermax F5 microplate reader (Molecular Devices, Sunnyvale, CA). The amount of GFP for each sample is represented by the change in fluorescence units (ΔFU), or the mean fluorescence units (FU) of the infected sample over the mean FU of the uninfected sample.
All available baculovirus genomes from GenBank (as of September 2014) were uploaded to the CLC Bio Main Workbench. These genomes were manually searched for FP25K/open reading frame 61 (ORF61) homologues. The amino acid sequences of these FP25K/ORF61 homologues from 61 baculovirus genomes were extracted, compiled, and aligned by using CLC Bio Main Workbench.
The amino acid sequences of wild-type FP25K and both the D22P and L36P mutants were entered into the COILS server (http://www.ch.embnet.org/software/COILS_form.html), and the coiled-coil structure of these proteins was predicted (37). The output for this prediction is represented by graphical plots, with all residue-scanning windows (14, 21, 28) containing the probability of a coiled-coil structure based on similarity to structures in a database of known coiled-coil proteins. High-probability peaks in these plots indicate the likelihood of a coiled-coil structure in that region.
Inverse PCR and site-directed mutagenesis, as described above, were used to produce fp25k deletions and fp25k point mutations, respectively, in the template pFBHTa-AcFP-GFP(ProFP).
Inverse PCR was performed on pFBHTa-AcFP-GFP(ProFP) to produce ΔL13:L44, ΔG52:L139, and ΔK142:S206 deletions within FP25K. The following primers were used: AcFP-L13:L44-F and AcFP-L13:L44-R for ΔL13:L44, AcFP-G52:L139-F and AcFP-G52:L139-R for ΔG52:L139, and AcFP-K142:S206-F and AcFP-K142:S206-R for ΔK142:S206 (Table 1).
Site-directed mutagenesis was performed on pFBHTa-AcFP-GFP(ProFP) to produce D22P and L36P mutations within FP25K by using a QuikChange II XL site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA). The following primers were used: AcFP-L36P-F and AcFP-L36P-R for the L36P mutation and AcFP-D22P-F and AcFP-D22P-R for the D22P mutation (Table 1).
Both the deletions and point mutations produced were confirmed by sequencing. The resulting deletion and point mutation constructs were transformed into DH10Bac-fp::287 cells. The recombinant bacmid DNAs produced were used to transfect Tn5 cells, and the resulting BV was harvested.
The GenBank nucleotide sequence accession number for the AcMNPV fp25k and 287-bp Sf21 host DNA fusion sequences is JX569226.
Previous work on FP25K localization was largely inconclusive due to inconsistent results of fractionation and immunogold TEM studies (32). Therefore, we determined FP25K localization using a more quantitative strategy by the production of a virus with an AcMNPV FP25K-GFP fusion for quantitative measurement of GFP levels during infection of cells (Fig. 1B). In order to accomplish this, we first inactivated the endogenous fp25k copy of AcBacmid by passaging AcBacmid BV 10 times in Sf21 cells, followed by PCR screening (Fig. 1A, left). fp25k mutants with an inserted 287-bp piece of host cellular DNA (GenBank accession number KX284746) were prominently detected at passage 10 (Fig. 1A, left). As determined by Southern blot analysis of the viral genome using this 287-bp host DNA as a probe, this host DNA was inserted only once in the viral genome at the fp25k locus (data not shown) and resulted in the introduction of a premature stop codon in the 287-bp host DNA sequence. The AcBacmid DNA from passage 10 was then introduced into E. coli cells by transformation. A clone (C9) carrying the mutant fp25k gene and termed AcBac-fp25k::287 was isolated and confirmed by PCR (Fig. 1A, right). Next, the AcMNPV fp25k-gfp cassette was introduced into the polh locus of AcBac-fp25k::287 to create AcFP-GFP(ProPH). Fusion protein production was confirmed by immunoblotting of AcFP-GFP(ProPH)-infected Tn5 and Sf9 cells (Fig. 1C). Relative to the 27-kDa molecular mass of GFP from infection with the nonfused AcBacGFP control, it is clear that a full-length AcMNPV FP25K-GFP fusion protein was produced, which ran at the predicted fusion molecular mass of 51 kDa in both cell lines (AcMNPV FP25K, 24 kDa; GFP, 27 kDa). There was a small percentage of the AcMNPV FP25K-GFP fusion protein that was cleaved during infection of both cell lines. Previous studies also observed full-length 25-kDa and cleaved 23-kDa FP25K products from cell lysates of AcMNPV E2 infections, in addition to cleaved FP25K from Helicoverpa armigera nucleopolyhedrovirus (HearNPV) infections (16, 17). It was noted previously by Braunagel et al. (17) that the cleavage was likely caused by degradation during harvesting; however, we found that harvesting at 48 hpi yielded considerably fewer cleavage products than those at 72 hpi, suggesting that this process occurred during infection (Fig. 1C and data not shown).
Finally, in order to confirm the loss of function of the endogenous fp25k copy of AcBac-fp25k::287 and rescue by AcMNPV FP25K-GFP at the polh locus, we performed a BV production assay (Fig. 1D). Since AcBac-fp25k::287 is a mutant clone resulting from virus passage in cells and served as the viral background for the production of the fusion proteins in this study, it was important to test the loss of functionality by the mutant and complementation by the addition of the fusion protein. As infection progressed, particularly after late times (>24 hpi), the AcBac-fp25k::287 mutant produced more BV than did the AcBacGFP wild-type control, which has a functional fp25k gene, as previously reported (16, 18). At the same time, the AcMNPV FP25K-GFP fusion protein from AcFP-GFP(ProFP) complemented the mutant fp25k copy of AcBac-fp25k::287 since the level of BV production of AcFP-GFP(ProFP) was similar to that of the AcBacGFP control during cell infection. Since AcFP-GFP(ProFP) demonstrated normal levels of BV production, it was used for quantitative FP25K localization studies. Overall, these BV production assay results confirm the inactivation of endogenous fp25k in the AcBac-fp25k::287 mutant and the rescue of FP25K functionality in the virus producing the AcMNPV FP25K-GFP fusion [AcFP-GFP(ProFP)]. It is possible that the insertional mutant AcBac-fp25k::287 may encode a partial FP25K protein; nonetheless, this virus does not have full functionality in the context of BV production (Fig. 1D).
In order to investigate where AcMNPV FP25K was localized during infection, Tn5 and Sf9 cells were infected with AcFP-GFP(ProPH) and monitored by confocal microscopy at 12, 24, 36, and 48 hpi (Fig. 2A). A small quantity of AcMNPV FP25K-GFP was produced starting at 24 hpi and was found entirely in the cytoplasm. By 36 and 48 hpi, about 4% of the total FP25K was present in the nucleus, but it still accumulated predominantly in the cytoplasm, at 96% (Fig. 2B and andC).C). In addition, at these times, a large proportion of FP25K localized to the cytoplasm, and several GFP-labeled amorphous structures were observed in many cells (Fig. 2A, white arrowheads). Previously, these large FP25K amorphous structures were observed during AcMNPV infection (18). This localization appears to be FP25K specific, because the GFP localization was uniform across both cell lines (as seen for AcBacGFP control infection) (Fig. 2A). Also, the percentages of FP25K found in the nucleus and cytoplasm are almost identical between infections of Sf9 and Tn5 cells (Fig. 2B and andC).C). The same localization was also observed with an N-terminal GFP-AcFP25K fusion (data not shown) and an FP25K antibody (38).
In immunogold TEM studies, Harrison and Summers observed FP25K in electron-dense structures in the nuclear periphery, thought to be chromatin (32). We also found a small quantity of FP25K present in the nucleus (Fig. 2A). It is important to verify this finding because the role of and/or interactions with FP25K in the nucleus could be different from those in the cytoplasm. To confirm that AcMNPV FP25K-GFP was indeed found in the nucleus and was not from signal bleed-through, z-stack analysis was performed during AcFP-GFP(ProPH) infection of Tn5 and Sf9 cells (Fig. 2D). FP25K was found across multiple depths of the nucleus for both Tn5 and Sf9 cell infections. FP25K was found both in the center of the nucleus and near the nuclear envelope. Previously reported evidence supports the localization of membrane-associated FP25K to the INM surface because of its association with trafficking ODV envelope proteins from the outer nuclear membrane to the INM (26).
Overexpression of proteins can alter subcellular localization (39,–42). In AcFP-GFP(ProPH) infection, the AcMNPV FP25K-GFP fusion protein is under the control of the highly active polh promoter. Therefore, it was necessary to verify that overexpression of AcMNPV FP25K-GFP did not alter its localization within the cell. To achieve this, we produced the AcFP-GFP(ProFP) virus that expresses AcMNPV fp25k-gfp under the control of the native fp25k promoter (Fig. 3A). First, the promoter activities of AcFP-GFP(ProPH) and AcFP-GFP(ProFP) were compared by measuring the amount of GFP being produced during infection of Tn5 and Sf9 cells (Fig. 3B). During infection of both Tn5 and Sf9 cells, AcFP-GFP(ProPH) resulted in a >4-fold increase in GFP levels compared to those with AcFP-GFP(ProFP) infection. The subcellular localization of AcMNPV FP25K-GFP from infection of Tn5 cells with AcFP-GFP(ProPH) and AcFP-GFP(ProFP) was analyzed by confocal microscopy at 48 hpi (Fig. 3C). Considerably less overall FP25K-GFP and fewer and smaller amorphous structures were observed in AcFP-GFP(ProFP)-infected cells. However, as seen with AcFP-GFP(ProPH), some AcMNPV FP25K-GFP was also found to localize to the nucleus of AcFP-GFP(ProFP)-infected cells. Overall, the percentages of detectable FP25K in the nucleus (4% versus 5%) and cytoplasm (96% versus 95%) were very similar between the two viruses (Fig. 3D). These data support the notion that overexpression has no impact on the nuclear and cytoplasmic localization of FP25K in infected insect cells.
We observed localization of FP25K primarily to the cytoplasm. However, to date, the genetic determinant(s) of FP25K localization is uncertain. Braunagel et al. reported the association of FP25K with several ODV proteins and, using microsomal membranes, presented evidence suggesting that this may be occurring at the ER membrane (26). Protein domain searches indicated that AcMNPV FP25K is composed of a conserved N-terminal coiled coil, a putative actin binding helix, and a nucleic acid binding motif (43, 44). We hypothesized that because of its role in protein-protein interactions, the N-terminal coiled-coil domain of FP25K plays a role in its predominant cytoplasmic localization by associating with ODV proteins. To understand its conservation among baculoviruses, the FP25K or ORF61 homologue amino sequences were extracted from all available (as of 2014) baculovirus genomic sequences and aligned (Fig. 4A). By comparing the entire aligned sequences, it is evident that all alphabaculovirus FP25K sequences contain an N-terminal coiled-coil domain and a small fragment (5 to 52 aa in length) at the C-terminal end, both of which are lacking in the betabaculovirus FP25K sequences (Fig. 4A). This N-terminal coiled-coil domain sequence spanned from around alphabaculovirus amino acid position +5 to position +40 (or alignment positions +20 to +55). Among the alphabaculovirus FP25Ks, the N-terminal coiled-coil domain sequence shows a high level of conservation, with a score of 98 out of 100 from a T-Coffee alignment (Fig. 4B) (45).
To investigate whether this coiled-coil domain is necessary for protein localization in other related proteins, the AcMNPV FP25K amino acid sequence was run through a protein structure prediction server. Sequence-based similarity searches (e.g., BLAST) for FP25K have been largely unsuccessful in identifying proteins with similar sequences in order to predict the function(s) of FP25K and its corresponding domains. We employed a protein structure prediction search using the more sensitive HHpred program to identify more remotely related proteins (46). Almost the entire AcMNPV FP25K protein (aa 10 to 189 out of 214) displayed 99.9% structural similarity and a strong E value of 3.0E−23 with long interspersed element 1 (LINE-1) open reading frame 1 protein (ORF1p). FP25K from betabaculovirus PxGV also had 99.5% structural relatedness and an E value of 7.4e−14 with LINE-1 ORF1p. However, because PxGV FP25K does not contain the putative N-terminal coiled-coil domain, it aligned to only part of ORF1p. LINE-1 is an autonomous retrotransposon found in mammalian genomes, and its ORF1 protein functions as a nucleic acid chaperone (47). Like the alphabaculovirus FP25Ks, this ORF1p has a conserved N-terminal coiled-coil domain. ORF1p also contains a highly conserved noncanonical RNA recognition motif (RRM) and a basic carboxy-terminal domain (CTD). Goodier et al. found that the N-terminal one-third of ORF1p is necessary for cytoplasmic retention and protein localization (48). Interestingly, they also discovered that ORF1p coimmunoprecipitated and colocalized with several RNA binding proteins known to be part of cytoplasmic stress granules. This could offer some insight into the amorphous FP25K structures that we, and others, have observed in the cytoplasm of AcMNPV-infected cells (Fig. 2 and and3)3) (32). This demonstrates that this putative N-terminal coiled-coil domain present in alphabaculovirus FP25Ks could be involved in protein localization.
In order to address the role of the N-terminal coiled-coil domain in FP25K localization, a set of proline point mutations were made in the coiled-coil domain of the AcMNPV FP25K-GFP fusion protein at amino acid positions 22 and 36 (Fig. 5A, top). As predicted by the COILS program, a leucine or isoleucine at position 36 is highly conserved and critical for the formation of the coiled-coil structure, and the introduction of a proline here disrupted its predicted structure (Fig. 5A, bottom) (37). On the other hand, the aspartic acid at position 22 is variable and not required for the coiled-coiled structure, so a proline substitution here should not and did not impact its structure (Fig. 5A, bottom). Localization of AcFP-GFP(ProFP) (control), the L36P mutant (coiled-coil mutant), and the D22P mutant (control mutant) during infection of Tn5 cells was monitored by confocal microscopy at 48 hpi (Fig. 5B). More amorphous structures and overall GFP were observed in the nuclei of L36P mutant-infected cells than in the nuclei of AcFP-GFP(ProFP)- and D22P mutant-infected cells (Fig. 5B, white arrowheads). This was confirmed by quantitative analysis, where ~18% of GFP from L36P mutant infection localized to the nucleus, resulting in a >5-fold increase compared to levels from AcFP-GFP(ProFP) and D22P mutant infections (Fig. 5C).
Since these point mutations had only a moderate effect on phenotype, we decided to confirm the role of the coiled-coil domain in localization by deleting portions of the FP25K protein (Fig. 5D). A series of three deletions in FP25K were made: ΔL13:L44 (coiled-coil domain deletion), ΔG52:L139 (middle deletion), and ΔK142:S206 (C-terminal deletion) (Fig. 5D). Next, confocal microscopy was performed on Tn5 cells infected with these three deletion viruses, in addition to the full-length AcFP-GFP(ProFP) FP25K control (Fig. 5E). In contrast to the localization observed for the other deletions (ΔG52:L139 and ΔK142:S206) and the full-length control, it is clear that deletion of the N-terminal coiled-coil domain (ΔL13:L44) dramatically changed FP25K localization to a largely nuclear localization (Fig. 5E). Approximately 76% of GFP from ΔL13:L44 infection localized to the nucleus, whereas <4% of GFP was nuclear in AcFP-GFP(ProFP), ΔG52:L139, and ΔK142:S206 infections (Fig. 5F).
The difference in phenotype between ΔL13:L44 and the other viruses was more robust and apparent than that of the point mutations. Even though the L36P point mutation altered the predicted coiled-coil structure, it is possible that this FP25K point mutant could still maintain some weak association with ODV proteins or other partners in the cytoplasm. Also, it appears that even without the coiled-coil domain, some FP25K remained in the cytoplasm. However, we believe that this is attributed mainly to the time of infection at which microscopy was performed, because at later time points (>72 hpi), almost all of the FP25K-GFP made it to the nucleus during ΔL13:L44 infection (data not shown).
It appears that in addition to localization changes seen with the ΔL13:L44 virus, infections with the ΔG52:L139 and ΔK142:S206 mutants resulted in less FP25K amorphous structure formation (Fig. 5E). The strong punctate FP25K-GFP structures seen with wild-type AcFP-GFP(ProFP) infection appeared to be reduced considerably, leading to more homogeneous localization across the cytoplasm with the ΔG52:L139 and ΔK142:S206 mutants (Fig. 5E).
Previous studies showed that FP25K may play a key role in regulating the switch between and the ratio of the two virion phenotypes BV and ODV (18, 38). In fp25k mutants, fewer occlusion bodies and fewer ODVs per occlusion body are formed, but more BV is produced (16, 18, 23, 49). Therefore, to functionally investigate the role of the FP25K N-terminal coiled-coil domain in BV production, we performed a budding assay on the AcFP-GFP(ProFP), ΔL13:L44, ΔG52:L139, and ΔK142:S206 viruses (Fig. 6). Since the largest difference in BV production was observed at 72 hpi, as shown in Fig. 1D, this time point was represented in Fig. 6 and and8.8. As expected, and as shown in Fig. 1D, AcBac-fp25k::287-GFP (bacmid mutant) had 21- and 17-fold-higher levels of BV production than did AcBacGFP (wild-type bacmid) and AcFP-GFP(ProFP) (FP repair control), respectively. The coiled-coil ΔL13:L44 and C-terminal ΔK142:S206 viruses had 15- and 18-fold-higher BV titers than did AcFP-GFP(ProFP), respectively. However, the ΔG52:L139 virus did not show an increase in BV production. This finding suggests that the N-terminal coiled-coil domain and C-terminal region of FP25K are critical for the budding process of the virus.
The alignment of baculovirus FP25K sequences yielded an intriguing observation: none of the betabaculovirus FP25Ks contained an N-terminal coiled-coil domain (Fig. 4). It was reported previously that XecnGV FP25K can rescue some of the phenotypic defects of the BmNPV fp25k mutant, but it is uncertain how betabaculovirus FP25K localizes during infection (19). Based on our findings that the N-terminal coiled-coil domain determines the fate of FP25K in the cytoplasm, we hypothesized that since betabaculovirus FP25Ks do not contain this coiled-coil domain, PxGV FP25K would localize to the nucleus. To test this hypothesis, a PxGV FP25K-GFP fusion protein was produced and tracked during infection (Fig. 7A). First, the PxGV fp25k-gfp fusion cassette was introduced into the polh locus of AcBac-fp25k::287 C9, and fusion protein production was confirmed by immunoblotting of PxFP-GFP(ProPH)-infected Tn5 cells (Fig. 7B). Relative to the 27-kDa nonfused GFP product from AcBacGFP control infection, there was a full-length PxGV FP25K-GFP fusion protein being produced at the predicted molecular mass of 42 kDa (PxGV FP25K, 15 kDa; GFP, 27 kDa). As shown in Fig. 1C, there are some cleavage products detected, which were reported previously (17). This cleavage between FP25K and GFP may represent a conserved process across baculovirus FP25Ks, but more studies need to be done in order to confirm this result. The subcellular localization of PxGV FP25K was resolved by performing confocal microscopy on Tn5 cells infected with AcFP-GFP(ProPH) and PxFP-GFP(ProPH) viruses at 48 hpi (Fig. 7C). There was a stark contrast between the primarily cytoplasmic localization of AcMNPV FP25K-GFP and the more prominent nuclear localization of PxGV FP25K-GFP in AcFP-GFP(ProPH)- and PxFP-GFP(ProPH)-infected cells, respectively. In PxFP-GFP(ProPH) infection, 61% of PxGV FP25K-GFP localized to the nucleus, which was 17-fold higher than the amount of nuclear AcMNPV FP25K-GFP (Fig. 7D). This fold difference in nuclear localization was similar to that detected between ΔL13:L44 virus (coiled-coil deletion)- and AcFP-GFP(ProFP) control-infected cells (Fig. 5F). PxGV FP25K-GFP also localized primarily to the nucleus at 24 hpi (data not shown), which indicates that betabaculovirus FP25K localizes to the nucleus soon after expression and likely before nuclear membrane breakdown (3, 12, 13). In addition, it appears that FP25K amorphous structure formation was sustained during PxFP-GFP(ProPH) infection, as shown by the punctate FP25K structures observed in both the cytoplasm and nucleus (Fig. 7C, white arrowheads).
To verify that sequence and, more specifically, nuclear localization signal (NLS) differences were not the primary drivers of localization changes between AcMNPV FP25K and PxGV FP25K and that the presence or absence of the coiled-coil domain was the main contributor, the PxGV and AcMNPV FP25K sequences were analyzed for a putative NLS. According to NLS Mapper (a program that predicts nuclear localization signals), both PxGV FP25K and AcMNPV FP25K sequences have similar scores, indicating the presence of NLSs in both proteins (50). Therefore, it can be concluded that betabaculovirus FP25Ks localize to the nucleus because of their lack of coiled-coil domain structure and the presence of NLSs. The NLS present in alphabaculovirus FP25Ks does not drive its cytoplasmic localization, because it is likely hidden or masked by the structure or protein-protein interactions guided by the coiled-coil domain.
When BV production of the deletion viruses was assessed (Fig. 6), the resulting levels of BV from infections with the ΔL13:L44 (coiled-coil deletion) and ΔK142:S206 (C-terminal deletion) viruses were similar to those from infections with the fp25k mutant. In order to confirm that the N-terminal coiled-coil domain of alphabaculovirus FP25K plays a role in BV production, we determined the level of BV production of PxFP-GFP(ProFP) in Tn5 cells (Fig. 8). As with AcFP-GFP(ProFP), the PxFP-GFP(ProFP) virus expresses PxGV fp25k-gfp under the control of its native fp25k promoter. PxFP-GFP(ProFP) infection yielded 7-fold-higher BV titers than those of AcFP-GFP(ProFP) infection. The amount of PxFP-GFP(ProFP) BV (which contains PxGV fp25k-gfp in the polh locus of the fp25k::287 bacmid) produced approached and was not significantly different from that of the AcBac-fp25k::287 mutant. Therefore, PxGV FP25K is unable to rescue BV production to the wild-type level. This provides further evidence that the N-terminal coiled-coil domain of alphabaculovirus FP25Ks plays an important role in BV production and that betabaculovirus FP25Ks may not require this domain for the process of budding in betabaculoviruses.
The FP (few-polyhedra) phenotype was first observed over 40 years ago and is characterized by fewer OBs, a decrease in virion occlusion, and an increase in BV production, which provides a selective advantage for this phenotype in cell culture (18, 23, 25, 51, 52). This FP phenotype has been observed in multiple group I and II alphabaculoviruses and is often associated with mutations in the fp25k gene (16, 18, 23, 49, 51, 53, 54). Two localization studies on AcMNPV FP25K yielded disparate results (17, 32). By way of immunogold electron microscopy of AcMNPV-infected cells, Harrison and Summers found that FP25K localized to large cytoplasmic amorphous structures and electron-dense structures in the nucleus, thought to be chromatin (32). Contrary to these results, Braunagel et al. observed the presence of FP25K only in the cytoplasm (17). Using a virus that produced a functional AcMNPV FP25K-GFP fusion protein during infection (Fig. 1D), we detected 96% of AcMNPV FP25K in the cytoplasm, present mostly in large amorphous structures; however, 4% localized to the nucleus (Fig. 2A to toC).C). Nuclear AcMNPV FP25K was confirmed by z-stack analysis and was localized to both the center and the periphery of the nucleus (Fig. 2A and andD).D). Despite being under the control of the strong polh promoter, the expression levels of the FP25K-GFP fusion protein did not change its localization (Fig. 3).
Our localization data led us to ask, what is the genetic determinant(s) of FP25K localization? Braunagel et al. reported that membrane-associated FP25K, along with BV/ODV-E26 and importin-α-16, traffic and sort viral proteins across the INM (26). Several ODV envelope proteins associate with FP25K and/or BV/ODV-E26, such as ODV-E66 and ODV-E25 (26). From domain prediction, it is apparent that AcMNPV FP25K has conserved N-terminal coiled-coil, putative actin binding helix, and nucleic acid binding domains. Additionally, a protein structure prediction search using the AcMNPV FP25K amino acid sequence in HHpred resulted in an extremely high level of structural relatedness (99.9%) to LINE-1 ORF1p, which also contains an N-terminal coiled-coil domain previously shown to be necessary for cytoplasmic retention (46, 48). Based on its established function in protein-protein interactions and structural similarity to the domain in LINE-1 ORF1p, we hypothesized that the N-terminal coiled-coil domain of AcMNPV FP25K is responsible for its prominent cytoplasmic localization. When baculovirus FP25K amino acid sequences are aligned, the N-terminal coiled-coil domain is found in all alphabaculoviruses but none of the betabaculoviruses (Fig. 4). Both the coiled-coil domain point mutation to disrupt structure (L36P) and a complete deletion (ΔL13:L44) resulted in a shift of AcMNPV FP25K localization from the cytoplasm to the nucleus (Fig. 5). In addition, when PxGV FP25K localization was tested, it was found primarily in the nucleus (Fig. 7). This suggests that the N-terminal coiled-coil domain present in all alphabaculovirus FP25K sequences plays a key role in their localization to the cytoplasm, possibly by way of protein-protein interactions with ODV envelope proteins, as previously reported for AcMNPV FP25K (17). Additionally, many of these ODV envelope proteins contain N-terminal domains abundant in hydrophobic residues that are critical for FP25K or ODV-E26 interactions, and hydrophobic domains often associate with coiled-coil domains (55, 56). Based on its lack of an N-terminal coiled-coil domain and nuclear barrier during betabaculovirus occlusion, we propose that the roles of interaction and trafficking of ODV envelope proteins within the betabaculovirus life cycle may not be performed by their FP25K homologues (3, 12, 13).
Coiled-coil regions within cellular and viral proteins are critical for protein localization, such as the Golgi protein giantin (57), the T-cell activation scaffold protein CARMA1 (58), the transcription factor promyelocytic leukemia retinoic acid receptor alpha (PML-RARα) (59), foamy virus Gag protein (60), and the integral membrane glycoprotein tetherin (61). At times, the coiled-coil domain function in subcellular localization is associated with homo- or hetero-oligomerization of these proteins. Another prominent example of these properties is LINE-1 ORF1p, which has strong structural relatedness to the FP25K orthologues in alphabaculoviruses and betabaculoviruses. LINE-1 ORF1p contains an N-terminal coiled-coil domain, a noncanonical RRM, and a basic CTD, all of which are highly conserved (47). These predicted features are found in alphabaculovirus FP25Ks and, with the exception of the N-terminal coiled-coil domain, betabaculovirus FP25Ks. Like the alphabaculovirus FP25Ks, the N-terminal coiled-coil domain of ORF1p was also found to be a critical determinant of its cytoplasmic retention and protein localization (48). Additionally, the N-terminal coiled-coil domain of ORF1p is also important for its oligomerization, which is necessary for ORF1p RNA binding activity (47). Not only did ORF1p localize to the cytoplasm, it also was present near large cytoplasmic foci identified as stress granules. Intriguingly, we, and others, observed that the alphabaculovirus FP25K protein studied also localized to large cytoplasmic amorphous structures (32). Stress granules are cytoplasmic aggregates assembled in response to stress conditions like heat shock, oxidative stress, and viral infection and are necessary for cellular maintenance of gene expression, homeostasis, and cytopathology (62). As with other cellular processes, these stress granules can be controlled and manipulated by viruses to promote viral replication, viral translation, and antiviral evasion. Recently, Kotani et al. found that the BmNPV nucleic acid binding protein baculovirus repeated open reading frames B/E (BRO-B/E) localized to amorphous cytoplasmic foci and associated with a cellular translation regulator, T-cell intracellular antigen 1 homologue (BmTRN-1), a known component of stress granules (63). They also determined that BRO-B/E expression led to decreased protein synthesis and proposed that this may be an additional mechanism for controlling protein synthesis in baculoviruses. Since alphabaculovirus FP25Ks appear to localize to similar amorphous cytoplasmic structures with or without infection (data not shown) (structures are smaller in uninfected cells), and regardless of the expression level, we believe that these structures are part of the virus life cycle and are controlled by baculoviruses for their benefit through posttranscriptional regulation, perhaps in part through FP25K. Further studies are needed to investigate these FP25K cytoplasmic foci and their potential identity as stress granules along with the predicted activity of nucleic acid binding in FP25K and possibly FP25K oligomerization involved in this process.
The phenomenon of higher-level BV production, less virus occlusion, and OB production in alphabaculovirus fp25k mutants is well established in the literature (8, 16, 18, 23, 24, 49, 54). We looked at BV production by the AcMNPV FP25K deletion viruses (ΔL13:L44, ΔG52:L139, and ΔK142:S206) and found that the coiled-coil domain (ΔL13:L44) and C-terminal domain (ΔK142:S206) deletions resulted in higher levels of BV production, reaching levels similar to those of the fp25k mutant (AcBac-fp25k::287) (Fig. 6). Since the FP25K C-terminal deletion (ΔK142:S206) uncouples localization changes and increased BV production, it would be intriguing to investigate its role in FP25K function. The PxFP-GFP(ProFP) virus also increased levels of BV production to nearly fp25k mutant levels (Fig. 8). This is slightly dissimilar from the BV production results reported by Nakanishi et al., who used the BmNPV system expressing fp25k from various baculoviruses, including XecnGV (19). In that study, those authors observed that XecnGV FP25K could partially restrict BV production to wild-type BmNPV T3 levels. However, some of the results from their functional assays of BmNPV-XecnGV FP25K were different from those reported previously, including ODV occlusion and V-cathepsin (V-CATH) activation. They attributed this to XecnGV fp25k expression level disparities from the alphabaculovirus counterparts, which may be the cause of the variance with our budding results. Overall, our BV production assay results indicate that the N-terminal coiled-coil domain of alphabaculovirus FP25Ks plays a negative role in BV production, possibly through interaction with and trafficking of ODV envelope proteins. Previous studies by Braunagel et al. and Hong et al. suggest that insertion of ODV envelope proteins into the INM, possibly by reaching a critical mass, is needed before blebbing can occur and intranuclear microvesicle precursors can be observed (26, 55). This N-terminal coiled-coil domain may be critical for the efficient transport of ODV envelope proteins to the INM, blebbing, and eventual formation of ODV to occur. Therefore, without this domain, the switch from BV to ODV is delayed, leading to high-level BV production.
Alphabaculovirus FP25Ks are thought to be partially responsible for the general switch between the production of BV and the production of ODV (14,–16). In addition, infection with alphabaculovirus fp25k deletion mutants leads to the accumulation of BV envelope proteins like GP64 (17). Certain BV envelope proteins, namely, GP64 or F proteins, are required for the process of virion budding to occur (64). Accumulation of GP64 at cell membranes has been shown to be necessary for budding. Also, recently, Li et al. reported that more GP64 was incorporated in AcMNPV fp25k mutant BV than in wild-type BV (29). This led to higher BV infectivity with the fp25k mutant. It may be that without a functional alphabaculovirus FP25K protein, ODV membrane proteins rely on transportation from the other sorting proteins, E26 and importin-α-16, and are not optimally trafficked to the INM for robust occlusion. Therefore, the switch to ODV production is delayed, and BV envelope proteins constitutively accumulate in the envelope, leading to higher levels of BV production. Moreover, it has also been thought that the accumulation of ODV envelope proteins like E25 in the nucleus is necessary for the retention of nucleocapsids in the nucleus for occlusion (4). Less-than-optimal trafficking of these ODV envelope proteins in fp25k mutants could greatly delay this retention step, leading to higher levels of nucleocapsid egress and subsequent budding. The N-terminal region of alphabaculovirus FP25K was shown to be critical for the interaction of ODV-E66 and -E25 with FP25K (17). Our data showing higher levels of BV production from the coiled-coil deletion maintain this interaction as an important step in making the switch from BV production to occlusion (Fig. 6), possibly through delayed or suboptimal trafficking of these proteins to the nucleus and delayed nucleocapsid retention. Additionally, betabaculovirus FP25Ks may not perform this trafficking function because of their lack of an N-terminal coiled-coil domain, thereby also leading to higher levels of BV production (Fig. 8).
Previous phylogenetic studies of baculoviruses suggested that they coevolved with their insect hosts (65, 66). From the perspective of fp25k gene phylogeny and function, alphabaculoviruses evolved to form and occlude into polyhedra dissimilarly from betabaculovirus granules, thereby requiring an additional function of trafficking of ODV membrane proteins. Furthermore, this occlusion trafficking function was not needed during betabaculovirus infection due to the fact that the nuclear barrier is broken down before betabaculovirus occlusion occurs; however, as alphabaculoviruses evolved to keep the nucleus intact during the occlusion phase, this function became necessary. As a result, the alphabaculovirus fp25k sequence diverged and gained the function of the N-terminal coiled-coil domain, unlike its betabaculovirus predecessor. The mechanism of this speciation could be explained by 3 likely hypotheses: (i) transposition of host DNA into the fp25k locus during infection, (ii) a recombination event during an ancestral alphabaculovirus coinfection with another virus that contains a coiled-coil domain, or (iii) a domain duplication event.
The data that we present indicate that AcMNPV FP25K localizes primarily to the cytoplasm in large amorphous structures; however, as infection progresses, there is a small quantity of FP25K that accumulates in the nucleus. It is possible that this nuclear FP25K is functionally different from the cytoplasmic form. However, more studies are needed to investigate this observation. Braunagel et al. detected FP25K in purified BV and ODV, and in addition, Harrison and Summers observed immunogold-labeled FP25K bound to nuclear chromatin structures; therefore, looking at nuclear FP25K colocalization or proximity to nucleocapsids or replication centers may be helpful (17, 32). Based on its disparate localizations between the alpha- and betabaculovirus genera, the role of FP25K appears to be distinctive in each life cycle. The N-terminal coiled-coil domain of alphabaculovirus FP25K genetically sets it apart from betabaculovirus FP25K species and is an important determinant of its localization. Deletion of this domain, in addition to parts of the C terminus, appears to alter BV production by alphabaculovirus species. It is likely that without the N-terminal coiled-coil domain, betabaculovirus FP25K is not involved in trafficking like its alphabaculovirus counterpart. Due to the fact that betabaculovirus FP25K is found in the nucleus early during infection, and as soon as it is produced, it likely has a function in this organelle before nuclear membrane breakdown. In order to elucidate the functionality of betabaculovirus FP25K, a betabaculovirus bacmid system would be paramount in this endeavor.
We thank both Andor Kiss and Richard Edelmann from the Center for Bioinformatic and Functional Genomics and the Center for Advanced Microscopy and Imaging, respectively, at Miami University for their technical assistance. We thank J. L. Xue for providing pp34 primer information. We also thank Eileen Bridge for proofreading of and suggestions for the manuscript.
This work was supported by the U.S. Department of Agriculture (U.S.-Egypt Science and Technology Joint Fund project no. 58-3148-7-164) and the Miami University Interdisciplinary Research Round Table Fund Project.