INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Autographa californica nucleopolyhedrovirus (AcMNPV) few polyhedra (FP) mutants were first observed by Hink and Vail in infected Trichoplusia ni cultures (16). By definition, FP mutants are viral isolates which produce few (<10) polyhedra compared to wild-type (WT) viral isolates, which produce 50 or more polyhedra per cell. Mutations in several regions of the AcMNPV genome can result in an FP phenotype. Examples include insertion of the copia-like transposable element (TE-D) in the HindIII-K region of the baculovirus genome (map units [m.u.] 85.1 to 87.2 [24]), deletions within the PstI-G fragment (m.u. 8.6 to 10.2 [21]), deletions within the PstI-I fragment (m.u. 14.3 to 17.9 [21]), and mutations within the FP25K gene. Using nine FP mutants and marker rescue, Fraser et al. (12) identified a region of the genome that was consistently altered in an FP phenotype (HindIII-I region, m.u. 33.8 to 37.7) and noted that the FP virus-infected cells were missing a 25-kDa protein. The coding region for this 25-kDa protein, called the FP25K gene, was sequenced by Beames and Summers (2), and FP25K gene sequences are now available for five baculovirus species (Fig. 1). This study addresses only features of an AcMNPV FP phenotype due to mutations within the FP25K gene.

View larger version (70K): [in this window] [in a new window]   FIG. 1.   Genomic constructs and sequence comparison of FP25K proteins. (A) The 5' untranslated region of FP25K (E2) is shown, along with the TAAG transcription initiation sequence (arrow). In the 480-1 mutant virus 120 nt have been removed, as indicated by dashes. (B) The amino acid sequences of five baculovirus FP25K protein sequences are compared. Identical amino acids are shaded and outlined, while conservative changes are shown in the shaded regions. The 480-1 N-terminal methionine is identified (arrow), as is the site of the -gal fusion (arrow). The N-terminal region of FP25K contains a conserved coiled-coil domain (underlined and marked with asterisks), and the C-terminal region contains a putative actin binding helix (underlined; the asterisk shows the requisite E or Q). Rules used to assign conservation are as follows: A = G = S = T, V = L = I = M = F = Y = W, N = Q = D = E, and R = K = H. Accession numbers: BmMNPV (B. mori nucleopolyhedrovirus), L33180 (nt 43656 to 44298); GmMNPV (G. melonella nucleopolyhedrovirus), M29140; OpMNPV (Orgyia pseudosugata nucleopolyhedrovirus), U75930 (nt 50697 to 51323); AcMNPV, L22858 (nt 48513 to 49155); and LdMNPV (L. dispar nucleopolyhedrovirus), U58676.

The plaque phenotype caused by mutations within the FP25K gene is a decreased number of nuclear occlusions; however, upon more careful examination, the effects of FP25K gene mutations are more complex. The number of occlusions produced by an FP mutant can vary according to cell type. In vitro, FP mutant infections produce fewer occlusions in the T. ni cell line (TN-368) than in Spodoptera frugiperda cells lines (Sf21 or Sf9) (reference 11 and unpublished observations), and when larvae are infected by hemocoelic injection with an FP mutant, the number of occlusions produced per cell varies according to the type of infected tissue (27). In addition, FP mutant occlusions contain fewer virions than WT occlusions, predominantly of the enveloped single nucleocapsid type, and the envelopment of the occluded baculovirus form (occluded derived virus [ODV]) is impaired (10, 23, 25, 27). The altered and decreased ODV envelopment may correlate with the observed altered morphology and electron density of virus-induced intranuclear membranes (13). Cells infected with FP mutants release more budded virus (BV) into the medium (10, 13, 26), and virus progeny are still observed budding at the plasma membrane at 72 h post infection (p.i.) (26), whereas in cells infected with WT virus, BV production has decreased to barely detectable levels by 72 h p.i. (33). The prolonged period of BV cell surface maturation is reflected in increased BV titers. Potter et al. (26) showed that FP mutants of T. ni nucleopolyhedrovirus increased in frequency in vitro from undetectable levels up to 93% of total plaques in 10 serial passages, while Fraser and Hink (10) demonstrated that the FP phenotype in Galleria melonella NPV increased from undetectable levels up to nearly 100% of the cells showing an FP phenotype in five serial passages. The Lymantria dispar NPV produces an FP phenotype even more rapidly, with several isolates showing a 92% plaque FP phenotype in only one passage (30). Thus, the FP phenotype provides a selection advantage for in vitro virus maturation. However, there is no observed selection for the FP phenotype when virus is passaged by per os feeding of insect larvae (10). Hence, the FP phenotype resulting in reduced ODV and enhanced BV maturation is not an advantage for horizontal transmission from insect to insect in the natural infection of insect populations.

In an FP mutant-infected cell the rate of polyhedrin mRNA expression is significantly reduced compared to that of WT infection (15), and polyhedrin localizes less efficiently to the nucleus during the early occlusion phase of infection (24 h p.i.) (20), yet polyhedrin mRNA stability is similar in WT- and FP mutant-infected cells (15). While polyhedrin gene expression is significantly altered, the steady-state mRNA levels of another very late gene, the p10 gene, remain unaltered (15).

In summary, while the FP plaque phenotype is the most easily identified effect of mutations within the FP25K gene, such mutations also result in (i) increased production of BV, (ii) decreased amounts of viral occlusions, (iii) decreased amounts of ODV production, (iv) altered morphology of intranuclear membranes, (v) decreased amounts of polyhedrin mRNA, and (vi) altered transport of polyhedrin protein into the nucleus. However, little is known about the function(s) of the FP25K protein, even though immunoelectron microscopy detects FP25K protein in electron-dense regions in both the cytoplasm and nucleus (13). Considering the varied effects of mutations within the FP25K gene, it is possible that FP25K has more than one function during the invasion and infection process.

In this study we show that FP25K is a structural protein of the nucleocapsid of AcMNPV and that it may interact with or influence the levels or rate of protein accumulation of several structural proteins of the virus. Infection by two FP mutants, 480-1 and FP-gal, result in significantly decreased amounts of ODV-E66 protein; however, steady-state levels of ODV-E66 transcripts remain unchanged compared to the WT. Transport of ODV-E66 into the nucleus is impaired in both of these mutants, with the 480-1 mutant having the most significant effect. Transport of ODV-E25 (28) into the nucleus is different for the two mutants: ODV-E25 transport is delayed during infection by 480-1 but is unaffected when cells are infected with FP-gal. Studies designed to identify proteins interacting with FP25K suggest that ODV-E26 and ODV-E66 may interact directly with FP25K and that ODV-E25 and p39 may be components of a complex containing FP25K.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Insect cell lines and virus. S. frugiperda IPLB-Sf21-AE clonal isolate 9 (Sf9) cells were cultured in suspension at 27°C in TNMFH medium (31) supplemented with 10% fetal bovine serum (complete medium). AcMNPV (strain E2) was used to infect cells at a known multiplicity of infection (MOI), with time zero set at the time of virus addition. After 1 h of adsorption, cells were washed and resuspended in fresh complete medium. The FP mutant viruses 480-1 and FP-gal were described by Beames and Summers (1, 2) and are summarized in Fig. 1.

Western blot analysis of infected cells, virus, and virus fractionation. Sf9 cells were infected with either AcMNPV (WT) or 480-1 or FP-gal virus (MOI, 20), and at an appropriate time p.i., the cells were collected and washed once with phosphate-buffered saline (PBS). Cell pellets were resuspended in PBS containing protease inhibitors (20 µg of leupeptin per ml, 20 µg of aprotinin per ml, 20 µg of pepstatin A per ml, 0.5 mM phenylmethylsulfonyl fluoride, 1 µM E64). Cells were broken by sonication, and protein concentrations were determined by the method of Bradford (4).

Immunofluorescence microscopy. Cells were processed for light microscopy by using a modification of previously described procedures (8). Sf9 cells were infected (MOI, 20), and at the appropriate time p.i. cells were rinsed with Grace's medium and fixed with 3.7% paraformaldehyde in PBS (20 mM phosphate, 140 mM NaCl, pH 7.2) for 10 min at room temperature. The fixative was removed, and cells were washed and permeabilized with methanol (10 min) and subsequently treated with 0.5% Triton-X 100 (10 min), followed by two rinses with PBS. The cells were blocked for 1 h in 1% normal goat serum-3% bovine serum albumin and then incubated with primary antibody (anti-FP25K, no. 2804, 1:1,000; anti-E66, no. 5297, 1:1,000; or anti-E25, 1:2,000 in 1% normal goat serum in PBS) overnight at 4°C. Cells were rinsed three times, and secondary antibody (fluorescein isothiocyanate [FITC] (Sigma, St. Louis, Mo.; 1:100 in PBS) was added and incubated for 1 h. The cells were washed, and the nucleus was visualized by staining with DAPI (4',6-diamidino-2-phenylindole) (0.1 µg/ml in PBS). Cells were viewed and photographed with a Zeiss Axiovert 135 photomicroscope. Each experiment was performed three to five times, and observations were made by using both FITC- and tetramethyl rhodamine isothiocyanate-labeled secondary antibodies. Thus, several thousand cells were viewed per experiment, and representative cells were chosen for data presentation.

Quantitative primer extension. Sf9 cells were infected with either AcMNPV or FP-gal (MOI, 10), and infected-cell RNA was isolated by the method of Chirgwin et al. (9). Primer extensions were performed with 30 µg of RNA hybridized to specific oligonucleotide probes labeled with [-³²P]ATP (29). The oligonucleotide sequence of the ODV-E66 probe was 5'-GATAGGTACAAAAAACATATTAAAAATATTA CAACTATGAC-3', and that of the vp39 probe was 5'-CGCGAAAATGCAGCGATTAACTCTCATTTGTCGCGGCGCC-3'. RNA-primer hybrids were precipitated with ethanol, washed with 70% ethanol, and resuspended in 30 µl of reverse transcription reaction mix (50 mM Tris [pH 7.6], 60 mM KCl, 10 mM MgCl₂, 0.66 mM deoxynucleoside triphosphates, 1 mM dithiothreitol, 40 µl of RNasin, 50 µl of actinomycin D per ml, and 150 U of Moloney murine leukemia virus reverse transcriptase [U.S. Biochemicals, Cleveland, Ohio]). Extension of the annealed primers was performed at 42°C for 2 h. The reaction products were ethanol precipitated and resuspended in 2 µl of 0.1 N NaOH. After a 30-min incubation to eliminate the RNA template, 4 µl of sequencing stop buffer was added, and the samples were boiled for 3 min and analyzed by electrophoresis on a urea-6% polyacrylamide gel together with a sequencing ladder generated by using the same oligonucleotides. The gels were dried and subjected to autoradiography, and the primer extension products were quantitated with the FUJIX BAS2000 bioimaging analyzer system (Fuji Photo Film Co., Tokyo, Japan).

Immunoprecipitation. A total of 1.5 × 10⁶ infected cells (MOI, 20) were used for each immunoprecipitation. At the appropriate time, cells were collected and resuspended in 500 µl of TBN buffer (50 mM Tris [pH 8.0], 150 mM NaCl, 10 µg of leupeptin per ml, 10 µg of aprotinin per ml, 1 µg of pepstatin A per ml, 1 mM phenylmethylsulfonyl fluoride, 1 mM E64) supplemented with either 0.2% Tween 20 or 1% Nonidet P-40 (NP-40). Cells were incubated for 20 min at 4°C and then lysed by passage through a 25-gauge needle four times. The lysed extract was centrifuged (3,000 rpm, 10 min, 4°C, Microfuge), and the supernatant was preabsorbed for 1 h with 25 µl of preimmune serum at 4°C. Protein A-agarose (Sigma) (40 µl of a 50% slurry) was added to the extract and incubated for 1 h at 4°C. The immune complexes formed during preadsorption were pelleted at 3,000 rpm for 15 min in a Microfuge. The preadsorbed extract was then immunoprecipitated with the appropriate antibody overnight (25 µl, 4°C) (ODV-E66, no. 5297; FP25K, no. 2804; gp67, AcV₁). Protein A-agarose (40 µl, 50% slurry) was added and incubated for 2 h at 4°C. The agarose beads were washed three times in TBN and then once in TBS. The washed beads (20 µl) were prepared for SDS-PAGE (4% stacking gel, 12.5% separating gel).

Yeast two-hybrid library construction and screen. Sf9 cells were infected (MOI, 20), and after 1 h of adsorption, the virus inoculum was removed and the cells were resuspended in complete medium. At 18 or 24 h p.i., cells were collected and mRNA was isolated by using either the Poly A Tract System 1000 (Promega, Madison, Wis.) or the Poly(A) Pure mRNA Isolation Kit (Ambion, Austin, Tex.). The cDNA library was then constructed by using the Two Hybrid cDNA Construction Kit (Clontech, Palo Alto, Calif.). The libraries were amplified, and the resulting titers of the amplified libraries were as follows: 18 h p.i., 1.31 × 10¹²; 24 h p.i., 3.25 × 10¹³. To harvest large quantities of DNA from each library, a 1-ml aliquot of amplified library was diluted and grown on 200 Luria broth-ampicillin supplemented plates (150-mm diameter), bacteria were harvested, and plasmid DNA was purified by using the Plasmid Giga Kit (Qiagen, Valencia, Calif.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

FP25K homology comparison and selection of mutant viruses. One goal of this study was to examine the effects of FP25K deletions and potential interactions between FP25K and structural proteins of AcMNPV. To consider a design for FP25K gene mutations, we examined the predicted structural features of the FP25K protein. Computer-assisted analysis revealed a highly conserved coiled-coil domain (a structural motif often involved in protein-protein interactions) at the N terminus and a putative actin binding helix (Fig. 1). Two previously reported viral mutants (1, 2), each lacking one of these regions, were selected for these studies. The 480-1 mutant virus contains a 120-nucleotide (nt) deletion extending from position 45 to +77 (relative to the FP25K initiation codon), resulting in translation initiation at an internal Met (Fig. 1A). Thus, in this mutant the putative coiled-coil domain was deleted, resulting in an FP25K protein with a 31-amino-acid N-terminal truncation (13) (Fig. 1). The FP-gal mutant lacks the C-terminal half of FP25K, including the putative acting binding helix, and instead contains the amino acids -gal. Note that although the FP25K protein is highly conserved throughout the putative actin binding region, the last 20 to 25 amino acid residues show little conservation (Fig. 1B).

Two forms of FP25K are present during infection, and FP25K is a structural protein of AcMNPV. Previous work using Western blot analysis detected trace amounts of FP25K in purified BV, ODV, and viral occlusions; however, the levels were too low to be convincing (14). To clarify this, we used the more sensitive detection techniques involving horseradish peroxidase-labeled secondary antibody and chemiluminescence. This confirmed that FP25K was a structural protein in the nucleocapsids of both BV and ODV and further identified two forms, of 25 and 23 kDa (Fig. 2B, lanes 1 to 6). We know from previous work that BV can be contaminated with nonoccluded ODV if purified late during infection (unpublished observations). To decrease this potential contamination, we purified the BV from the supernatant of a 36-h-p.i. infected sample and tested the BV for the presence of ODV envelope proteins (see Fig. 3F). No ODV proteins were detected with the BV. A time course analysis of infected cells showed that the 25-kDa protein was the first to accumulate to detectable levels, while the 23-kDa protein was detected, and accumulated, later during the infection (Fig. 2A, lanes 2 to 9). The molecular mass of the truncated protein produced from the 480-1 mutant virus was similar to that of the 23-kDa FP25K species observed in WT-infected cells (Fig. 2A, compare lanes 8 to 11). The molecular mass of the FP-gal fusion was high (>110 kDa), and only one major protein was detected throughout infection (Fig. 2A, lanes 13 and 14).

View larger version (31K): [in this window] [in a new window]   FIG. 2.   FP25K is a structural protein of baculovirus. (A) Time course of FP25K accumulation in E2-infected cells (lanes 2 to 9) and size and late accumulation in cells infected with 480-1 mutant virus (lanes 10 to 12) and FP-gal recombinant virus (lanes 13 and 14). (B) BV and ODV were purified and separated into envelope (E) and nucleocapsid (C) fractions (lanes 1 to 6). WT AcMNPV-infected cell extract (48 h p.i.) were used a positive control (lane 7). Samples were separated by SDS-PAGE, Western blotted onto a polyvinylidene difluoride membrane, and reacted with antibody to FP25K. The amount of protein analyzed per lane is indicated. Numbers on the left are molecular masses in kilodaltons.

Mutations within the FP25K gene alter structural protein profiles during infection. Harrison and Summers (13) observed aberrant envelope-nucleocapsid interactions within the nuclei of FP25K mutant-infected cells. Instead of circular, well-formed microvesicles, some intranuclear membranes are elongated, angular, and unusual in shape and electron density. Additionally, although nucleocapsid assembly is apparently normal, there are few ODVs (intranuclear enveloped nucleocapsids) present in an FP25K mutant-infected cell (9, 23, 27). These observations show that intranuclear membrane formation, ODV envelopment, and the process of viral occlusion formation are altered in an FP mutant infection and suggest that these maturation processes could be at least partially influenced by FP25K. We decided to determine if mutations within the FP25K gene resulted in detectable changes in the amount or localization of baculovirus structural proteins. In a matched experiment, Sf9 cells were infected with WT, 480-1, or FP-gal virus, and the cells were harvested at various times p.i., protein concentrations were determined, and test gels were stained with Coomassie blue to visually confirm equivalent protein amounts per lane (data not shown). Matched gels and Western blots were then prepared, bound to antibodies and exposed to X-ray film for chemiluminescent detection. The exposure time was set to be optimal for protein detection in WT infection. Thus, differences in blot intensities represent quantitative changes in protein amounts relative to those of WT virus (E2 strain). During WT infection, the FP25K protein (25 kDa) was first detected at 18 h p.i. (Fig. 3A). The second form of FP25K (23 kDa) was detected at 36 h p.i. and accumulated to high levels by 72 h p.i. (see Fig. 2 for the molecular masses of two forms of FP25K and the protein produced by 480-1 virus). The truncated protein produced by the 480-1 mutant virus (23 kDa) was not detected until 36 h p.i. and did not accumulate to significant levels until 48 h p.i. (note that this pattern of protein accumulation was very similar to the appearance of the 23-kDa form during WT infection). The temporal detection of the FP-gal fusion was similar to that of the WT protein (Fig. 3A).

View larger version (89K): [in this window] [in a new window]   FIG. 3.   Structural protein-temporal expression. (A to E) Matched sets of cell lysates (25 µg/lane) were collected at defined times p.i. and analyzed for protein content with the appropriate antibody. SDS-PAGE, Western blotting, and antibody reactions were performed for matched sets, and the optimal exposure was set for WT AcMNPV. (A) FP25K; (B) ODV envelope proteins; (C) BV envelope proteins; (D) capsid-associated proteins; (E) tegument-associated protein. (F) Purified BV from both WT AcMNPV- and 480-1-infected-cell supernatants. Virus was loaded at 10 µg/lane, and that for the 48-h-p.i. time point was loaded at 15 µg/lane. Antibodies to each of the structural proteins were used to determine protein composition. The star in panel B shows a higher background for ODV-56 against the 70-kDa cellular protein. The arrows in panel C show increased levels of the higher-molecular-mass form in FP mutant-infected cells.

Decreased levels of ODV-E66 protein are not due to decreased steady-state levels of ODV-E66 transcripts. Mutations in the FP25K gene result in decreased steady-state levels of polyhedrin gene transcripts (15). To determine if reduced ODV-E66 transcription could also explain the significant decrease in ODV-E66 protein, quantitative primer extension analysis of ODV-E66 was performed. As an internal control, primer extension was performed for vp39. Figure 4A shows the results for primer extension products of vp39 and ODV-E66. The placement of these products was determined by using a matched sequencing ladder (not shown). The extension products for each primer extension were quantitated by using a Bio-Imaging Analyzer, and the ratio of ODV-E66 to vp39 was determined. Figure 4B shows that in two different experiments at 24 and 48 h p.i., the ODV-E66/vp39 ratio remained unchanged for WT- and FP-gal-infected cells. These data suggest that the decreased amount of detectable ODV-E66 protein was not due to transcriptional down-regulation at the ODV-E66 locus. Quantitative transcription was not done for the 480-1 mutant, but we note that both mutants result in a significant decrease in ODV-E66 protein.

View larger version (42K): [in this window] [in a new window]   FIG. 4.   Quantitative primer extension analysis. (A) Quantitative primer extension of p39 and ODV-E66 from Sf9 cells infected with WT AcMNPV (E2) or FP-gal virus. Primer extensions were with a mock-infected RNA sample (lane M) and E2 and FP-gal RNA samples with primers for vp39 or ODV-E66. The locations of ODV-E66 and vp39 were established by using DNA sequencing ladders (not shown) and are indicated to the right and left of the autoradiograph. (B) Calculated ratios of ODV-E66 to vp39 RNA for two replicate sets of experiments at 24 and 48 h p.i.

Mutations within the FP25K gene alter the intranuclear trafficking of ODV-E66. Since steady-state levels of ODV-E66 transcripts were not altered in FP-gal mutant infection, possible mechanisms to explain the lack of detectable ODV-E66 protein include decreased ODV-E66 translation, altered ODV-E66 targeting or transport, and/or altered ODV-E66 protein stability. To test if either FP25K mutant virus resulted in abnormal transport of ODV-E66, matched infections of WT and FP mutant viruses were performed, and ODV-E66 localization was determined by immunofluorescence with rabbit antisera raised against ODV-E66 and FITC-conjugated antirabbit secondary antibodies. (Trying to show these data by using an overview of infected cells is difficult, so representative cells from these experiments were chosen for enlargement. An example of how we made this choice is shown in Fig. 7A and B. Since showing phase-contrast, FITC, DAPI, and dual exposures for every data set [which was performed for each experiment] would make the figures large, cumbersome, and difficult to present, only the FITC and FITC-DAPI dual exposures [with DAPI defining the area of the nucleus] are presented. One example of phase-contrast-FITC dual exposure is shown in Fig. 7g2 for reference. Additionally, control antibody reactions [uninfected and preimmune sera] were performed for every experiment; however, the background cross-reactivity was minimal and was reproduced as a black field. An example of such a control is shown in Fig. 7D.)

View larger version (42K): [in this window] [in a new window]   FIG. 5.   Cellular localization of ODV-E66. Sf9 cells were infected with WT AcMNPV, 480-1, or FP-gal virus. At the indicated times p.i., cells were collected, fixed, and reacted with antibody to ODV-E66 (no. 5297; 1:1,000) and FITC-conjugated secondary antibody. To visualize the nucleus, cells were also stained with DAPI, and a dual exposure of FITC and DAPI was performed. Dashed arrows show cytoplasmic labeling, and solid arrows show intranuclear labeling. (A, B, C, and D) FITC exposure; (a and b) dual exposure of DAPI and FITC labeling.

480-1 mutant infection delays, but does not inhibit, transport of ODV-E25 into the nucleus. While the overall amino acid sequences of ODV-E66 and ODV-E25 do not show significant homology, the amino acids at the N termini of ODV-E66 and ODV-E25 (23 and 24 amino acids, respectively) are similar and sufficient to target fusion reporter proteins to intranuclear microvesicles and ODV envelope (19). If these N-terminal sequences interact with FP25K, then altered transport of ODV-E25 into the nucleus during FP25K mutant infection might be expected. To test this, the nuclear localization of ODV-E25 in WT- and FP25K mutant-infected cells was determined. By 24 h p.i. ODV-E25 localized very efficiently to discrete regions in the nuclei of WT-infected cells (data not shown), and the intensity of localization and intranuclear fluorescence increased at 48 and 72 h p.i. (Fig. 6A and B, AcMNPV). In 480-1 mutant-infected cells, there was a significant delay of localization of ODV-E25 in the nucleus. By 48 h p.i. ODV-E25 was located predominantly in the cytoplasm (Fig. 6A, 480-1); however, by 72 h p.i., ODV-E25 was detected in the nucleus in a pattern similar to that for WT infection (Fig. 6B, 480-1). Like for WT infection, immunoelectron microscopy showed that at 72 h p.i. ODV-E25 was present in intranuclear microvesicles and the ODV envelope in the 480-1 mutant-infected cells (data not shown). The nuclear localization of ODV-E25 in the FP-gal mutant-infected cells was indistinguishable from that in WT infection: by 24 h p.i. ODV-E25 was detected predominantly within the nucleus (data not shown), and it remained intranuclear throughout infection (Fig. 6A and B, FP-gal).

View larger version (34K): [in this window] [in a new window]   FIG. 6.   Cellular localization of ODV-E25. Sf9 cells were infected with WT AcMNPV, 480-1, or FP-gal virus. At the indicated times p.i., cells were collected, fixed, and reacted with antibody to ODV-E25 (1:1,000) and FITC-conjugated secondary antibody. To visualize the nucleus, cells were also stained with DAPI, and a dual exposure of FITC and DAPI was performed. Dashed arrows show cytoplasmic labeling, and solid arrows show intranuclear labeling. (A and B) FITC exposure alone; (a and b) dual exposure of DAPI and FITC labeling.

Localization of FP25K in WT- and FP mutant-infected cells. To provide an overview of FP25K localization during infection, we used immunofluorescence microscopy of WT-, 480-1-, or FP-gal-infected cells (Fig. 7). Figure 7 also exemplifies how the representative cells shown throughout this work were selected. An example of these is shown in Fig. 7a1, a2, b1, and b2 and then enlarged in Fig. 7A and B. This analysis showed that significant amounts of FP25K were present in the cytoplasm throughout infection (Fig. 7A to C). It is known from previous work that FP25K protein is present in both cytoplasmic and nuclear fraction, and it accumulates in electron-dense regions (14). These discrete regions enriched in FP25K could be detected as the microscope operator "focused through" the cell; however, they were not easily discernible by using a single focus, as shown in Fig. 7. The mutant 480-1 FP25K protein was not detected until 48 h p.i. (Fig. 7E), and even then the levels were too low to convincingly localize this protein; however, most of the truncated protein appeared to be cytoplasmic. By 72 and 96 h p.i., the localization of the truncated protein more closely resembled that seen in WT infection (Fig. 7F and data not shown). The FP-gal fusion protein showed a pattern of localization similar to that in WT infection (Fig. 7G, g1, and g2). Since the results for FP-gal were so similar to those for the WT, only one time point is shown (48 h p.i.), and an example of the matched phase-contrast-FITC dual exposure is also shown (Fig. 7g2). Like the results seen with antibodies to ODV-E66 and ODV-E25, background FITC levels of FP25K antibody tested against uninfected cells were reproduced as black field (Fig. 7D).

View larger version (51K): [in this window] [in a new window]   FIG. 7.   Cellular localization of FP25K. (A to C) Cells were infected with WT AcMNPV. Panels a1, a2, b1, and b2 show an overview of the cell population (a1 and b1, FITC exposure; a2 and b2, FITC-DAPI dual exposure; the cells chosen as representative of the population are indicated (boxes) and then enlarged in panels A and B (FITC-DAPI dual exposure). (E and F) Cells infected with the 480-1 mutant virus (E and F, FITC-DAPI; e, FITC). (G) Cells infected with FP-gal recombinant virus (g1, FITC; G, FITC-DAPI; g2, FITC-phase-contrast dual exposure). The time p.i. for each set is indicated. (D) Background labeling of FP25K antibody against uninfected Sf9 cells. Dashed arrows show cytoplasmic labeling.

FP25K potentially interacts with other baculovirus structural proteins. The Western blot analysis showed that compared to WT infection, infections by the FP25K mutant viruses resulted in increased protein levels of both BV envelope proteins, BV/ODV-E26 and gp67 (Fig. 3C). Previous data show that FP25K is capable of interacting with BV/ODV-E26 (3) (summarized in Table 1). Antibody to gp67 (AcV₁) coprecipitated FP25K, and the total amount of precipitated FP25K decreased as the amount of gp67 decreased (Fig. 8A, lanes 2 to 5). When the reciprocal experiment was performed with antibody to FP25K, gp67 was not coprecipitated; however, the FP25K antibodies precipitated both the 23- and 25-kDa forms (Fig. 8A, lane 1).

View larger version (31K): [in this window] [in a new window]   FIG. 8.   Immunoprecipitation analyses. Sf9 cells were infected with WT AcMNPV and at the indicated times p.i., and cell lysates were prepared in the presence of either Tween 20 or NP-40. (A) Protein complexes were immunoprecipitated with antibody (Ab) to gp67 (lanes 2 to 5) or FP25K (lane 1). Lane 6, 30-h-p.i. cell lysate sample analyzed as positive control. (B) Protein complexes were immunoprecipitated with antibody to ODV-E66 and two different detergent conditions (0.2% Tween 20 [T20] or 1% NP-40). The immunoprecipitated protein complex (20 µl) was analyzed by SDS-PAGE, Western blotted to a polyvinylidene difluoride membrane, and reacted with antibodies to FP25K, p39, BV/ODV-E26, or ODV-E25. A 48-h-p.i. cell lysate was used as positive control (lane 1 for each sample).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The amino acid sequence of FP25K is highly conserved among baculoviruses. AcMNPV and Bombyx mori nucleopolyhedrovirus have 96.7% overall similarity, and a comparison of all the sequenced FP25K genes show that the only region lacking significant conservation is the last 19 to 26 C-terminal amino acids (Fig. 1). Immunoblot analyses performed with purified AcMNPV ODV and BV show that FP25K is a structural protein in the nucleocapsids of both progeny viruses (Fig. 2); however, immunofluorescence microscopy indicates that a large fraction of FP25K remains cytoplasmic throughout infection (Fig. 7). Two major forms of FP25K are observed during infection and in purified virus: 25 and 23 kDa. While the 23-kDa form may correspond to translation at an internal Met, further analyses will be required to characterize this second form and clarify its functional significance.

Computer analyses of the primary structure of FP25K predicted two regions of possible structural relevance: a coiled-coil region and a putative actin binding helix (Fig. 1). Viral mutants lacking either of these regions were chosen for further study, and infection by either FP25K mutant resulted in a decreased amount of detectable ODV-E66 (Fig. 3). Immunofluorescence microscopy showed that during infection by both mutants, ODV-E66 displayed an altered pattern of intranuclear localization. This effect was more dramatic during infection by the 480-1 mutant, with infection resulting in a cytoplasmic location of ODV-E66. In FP-gal-infected cells, some ODV-E66 was detected within the nucleus; however, this trafficking was delayed and diminished compared to that with WT infection (Fig. 5). Since the two mutants produced equivalent amounts of ODV-E66, the effect of the 480-1 mutation on nuclear transport of ODV-E66 is likely not related to a minimal protein requirement for initiation of nuclear transport. Additionally, while the amounts of ODV-E25 which accumulate during WT and 480-1 infection are similar, nuclear transport of ODV-E25 was delayed by approximately 48 h, while intranuclear transport was not affected, in cells infected with FP-gal. These results suggest that the N-terminal region of FP25K may be important for normal trafficking of ODV-E66 and ODV-E25. Lack of the N-terminal region does not result in detectable relocalization of ODV envelope proteins to the plasma membrane (data not shown) or incorporation into BV (ODV-E66, ODV-E56, ODV-EC27, and ODV-E25) (Fig. 3F). While the mutant FP-gal virus was generated from the parental E2 virus, 480-1 is a spontaneous, naturally occurring FP25K mutant virus, and it is possible that additional mutations are present (1).

Western blot analysis of a time course of infected cells showed that in addition to that of ODV-E66, accumulation levels of several baculovirus structural proteins were altered. Compared to those in WT-infected cells, the amounts of BV envelope proteins gp67 and BV/ODV-E26 and the major capsid protein p39 increased in mutant-infected cells. We note that the proteins which exhibited an altered temporal accumulation pattern during FP mutant infection (Fig. 3) were the same proteins that protein-protein interaction studies suggested interact with FP25K (Fig. 8 and Table 1). Results from immunoprecipitation studies suggest that FP25K may be a member of protein complexes which include gp67, ODV-E66, and BV/ODV-E26. While FP25K coprecipitates with each of these proteins, these proteins are not precipitated in the reciprocal experiment (with antibody to FP25K). This suggests that FP25K may be an integral component of these complexes, and when bound, FP25K is masked from antibody recognition. The interactions of FP25K and ODV-E66 and of FP25K and BV/ODV-E26 have been confirmed by using the complementary yeast two-hybrid technique (Table 1) (3). However, a yeast two-hybrid cross between FP25K and gp67 was negative, suggesting that this interaction may be mediated by other proteins. Only the 25-kDa form is detected coprecipitating with gp67, ODV-E66, or BV/ODV-E26 (Fig. 8) (3), while antibody to FP25K precipitates both the 25- and 23-kDa forms. These data suggest that these two forms may be functionally distinct.

The data presented here indicate that in addition to the other, better-characterized effects of mutations within the FP25K gene, the accumulation patterns of several structural proteins are altered, and intranuclear transport of both ODV-E66 and ODV-E25 is compromised. We note that in the absence of a fully functional FP25K gene, intranuclear localization of polyhedrin is also compromised during the early occlusion phase of infection (20). FP25K could be affecting protein accumulation and trafficking by regulating transcription, mRNA stability, translation, or altered protein stability of one or many viral proteins. While transcription of ODV-E66 was not altered during an infection with the FP-gal virus (Fig. 4), FP mutant virus infection can alter steady-state levels of polyhedrin (15). The phenotypic hallmarks of the FP mutant, including apparently normal numbers of nucleocapsids but lack of ODV envelopment and a decreased amount of viral occlusions, may not be due to the mutant FP25K protein per se but may result from improper stoichiometry, localization, or protein complex formation of other viral proteins. Studies aimed to further characterize the composition of protein complexes containing FP25K and their role in the trafficking pathway(s) of ODV envelope proteins are under way.