Baculovirus Per Os Infectivity Factor Complex: Components and Assembly

Entry of baculovirus into host insects is mediated by a per os infectivity factor (PIF) complex on the envelopes of occlusion-derived viruses (ODVs). Knowledge of the composition and structure of the PIF complex is fundamental to understanding its mode of action. By using multiple approaches, we determined the complete list of proteins (nine) in the PIF complex. In contrast to previous knowledge in the field, the core complex is revised to ∼230 kDa and consists of PIF1 to -3 but not PIF4. Interestingly, our results suggest that the PIF complex is formed in the cytoplasm prior to its transport to the nucleus and subsequent incorporation into ODVs. Only the full complex is resistant to proteolytic degradation in the insect midgut, implying the critical role of the entire complex. These findings provide the baseline for future studies on the ODV entry mechanism mediated by the multiprotein complex.

ABSTRACT

Baculovirus entry into insect midgut cells is dependent on a multiprotein complex of per os infectivity factors (PIFs) on the envelopes of occlusion-derived virions (ODVs). The structure and assembly of the PIF complex are largely unknown. To reveal the complete members of the complex, a combination of blue native polyacrylamide gel electrophoresis, liquid chromatography-tandem mass spectrometry, and Western blotting was conducted on three different baculoviruses. The results showed that the PIF complex has a molecular mass of ∼500 kDa and consists of nine PIFs, including a newly discovered member (PIF9). To decipher the assembly process, each pif gene was knocked out from the Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV) genome individually by use of synthetic baculovirus technology, and the impact on PIF complex formation was investigated. Deletion of pif8 resulted in the formation of an ∼400-kDa subcomplex. Deletion of pif0, -4, -6, -7, or -9 resulted in a subcomplex of ∼230 kDa, but deletion of pif1, -2, or -3 abolished formation of any complex. Taken together, our data identified a core complex of ∼230 kDa, consisting of PIF1, -2, and -3. This revised the previous knowledge that the core complex was about 170 kDa and contained PIF1 to -4. Analysis of the PIF complex in cellular fractions suggested that it is assembled in the cytoplasm before being transported to the nucleus and subsequently incorporated into the envelopes of ODVs. Only the full complex, not the subcomplex, is resistant to proteolytic attack, indicating the essentiality of correct complex assembly for oral infection.

IMPORTANCE Entry of baculovirus into host insects is mediated by a per os infectivity factor (PIF) complex on the envelopes of occlusion-derived viruses (ODVs). Knowledge of the composition and structure of the PIF complex is fundamental to understanding its mode of action. By using multiple approaches, we determined the complete list of proteins (nine) in the PIF complex. In contrast to previous knowledge in the field, the core complex is revised to ∼230 kDa and consists of PIF1 to -3 but not PIF4. Interestingly, our results suggest that the PIF complex is formed in the cytoplasm prior to its transport to the nucleus and subsequent incorporation into ODVs. Only the full complex is resistant to proteolytic degradation in the insect midgut, implying the critical role of the entire complex. These findings provide the baseline for future studies on the ODV entry mechanism mediated by the multiprotein complex.

INTRODUCTION

Baculoviruses are large, rod-shaped double-stranded DNA (dsDNA) viruses that infect insects from the families Lepidoptera, Hymenoptera, and Diptera. The Baculoviridae family contains four genera: Alphabaculovirus, Betabaculovirus, Gammabaculovirus, and Deltabaculovirus (1). Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV) is the type member of the alphabaculoviruses and the most well-studied baculovirus (2). There are two types of progeny viruses produced during a typical baculovirus life cycle, namely, the budded virus (BV) and the occlusion-derived virus (ODV). ODVs are embedded in occlusion bodies (OBs) and are responsible for oral infection within the insect midgut, while BVs cause systemic infection in other larval tissues.

In nature, baculovirus infection begins when OBs are ingested and dissolved in the highly alkaline and protease-rich midgut. The released ODV particles pass through the peritrophic membrane lining the gut and initiate infection in midgut epithelial cells. Successful oral infection depends on a group of viral proteins, called per os infectivity factors (PIFs), on the ODV envelope (3, 4). PIF0 (originally called P74 because the molecular weight is ∼74 kDa) was the first PIF found to be essential for oral infection. The deletion of PIF0 has no impact on infectious BV production but totally abolishes oral infectivity of ODVs (5). Eight additional PIFs were identified later and were named PIF1 (Ac119; ∼60 kDa), PIF2 (Ac22; ∼44 kDa), PIF3 (Ac115; ∼23 kDa), PIF4 (Ac96; ∼20 kDa), PIF5 (ODV-E56 or Ac148; ∼41 kDa), PIF6 (Ac68; ∼16 kDa), PIF7 (Ac110; ∼7 kDa), and PIF8 (Ac83; ∼96 kDa) (6–13).

All the PIF proteins are expressed in the late stage of virus infection, and they are envelope proteins of ODVs, which are assembled in the nuclei of the infected cells. Many PIFs contain the inner nuclear membrane sorting motif (INM-SM), which is believed to guide the synthesized polypeptides into the nucleus (14). Another common feature of PIFs is that all their genes are conserved in Baculoviridae. Homologs of pif genes are also present in a wide range of invertebrate large dsDNA viruses, such as white spot syndrome virus (family Nimaviridae) (15), salivary gland hypertrophy virus (family Hytrosaviridae) (16), Apis mellifera filamentous virus (17), nudivirus (family Nudiviridae) (18), and bracovirus (family Polydnaviridae) (19).

A stable PIF complex of >170 kDa was first identified which could withstand 2% SDS–5% β-mercaptoethanol and heating at 50°C for 5 min. When this partial denaturation method was applied to pif-knockout viruses, it was demonstrated that the complex contained PIF1, -2, and -3 as a stable core, while PIF0 was loosely associated with the core (20). Later on, by use of blue native PAGE (BN-PAGE), a preferred method for isolating multiprotein complexes in a native state (21), the PIF complex was found to be ∼480 kDa and to contain PIF0, PIF1 to -4, and PIF8, while a stable core complex of ∼170 kDa, containing PIF1, -2, -3, and -4, was revealed by the partial denaturation method (22). A proteomic study showed that PIF6, Ac5, and Ac108 were associated with PIF1 (22). Later, PIF6 was confirmed to be a component of the PIF complex (23), while Ac5 was not (24). Recently, PIF7 was also identified as a component of the complex (3). However, it remains unknown if the current list of complex members is complete. All PIFs except PIF5 are associated with the PIF complex, which highlights the pivotal role of the complex for oral baculovirus infection (25). The intact complex was required to protect PIFs from proteolytic degradation by OB endogenous proteinases during ODV release from OBs (23). Understanding the structure and function of this unique complex calls for a detailed study, particularly in light of the ubiquitous nature of pif genes in a wide range of invertebrate large dsDNA viruses.

The mechanism of PIF complex assembly is still largely unknown. Study of the core complex should help us to understand the structure and formation of the entire complex. However, so far, the published information on the core complex has not been consistent, especially regarding the role of PIF4. While PIF1, -2, and -3 are essential for core complex formation, deletion of pif4 did not completely impair the stable core complex but resulted in a smaller stable complex of ∼150 kDa (22). This ∼150-kDa complex was later found to be sensitive to proteolytic degradation, and the apparent inconsistency is due to the fact that the partial denaturation method used did not completely inhibit proteolytic degradation by the OB’s endogenous proteinase (23).

In this study, we systematically investigated the composition, assembly, and function of the baculovirus PIF complex to understand its role in virus entry. Multiple approaches, including BN-PAGE, liquid chromatography-tandem mass spectrometry (LC-MS/MS), and Western blotting, were used to obtain a complete list of components of the PIF complex. The impact of each component on the formation of the AcMNPV complex under natural conditions was then analyzed by BN-PAGE. We also addressed the initial intracellular assembly of the complex. Our results provide a likely complete list of PIF complex components, revise the previous understanding of the core complex, and shed light on the process of complex assembly and intracellular transport.

RESULTS

Disclosure of components of the PIF complex.

To date, all the known PIFs have been identified individually, but it is still not clear if there are hitherto unidentified members of the PIF complex. To reveal all components of the PIF complex, we chose three baculoviruses, AcMNPV, Bombyx mori NPV (BmNPV), and Helicoverpa armigera NPV (HearNPV), and analyzed the composition of their PIF complexes systematically. An integrative investigation by BN-PAGE, LC-MS/MS, and Western blot analysis was conducted (Fig. 1). Specific antibodies were raised against individual PIFs of AcMNPV and HearNPV to signal their presence in their native PIF complexes. Because PIF4, -6, and -7 failed to yield effective antibodies, AcMNPV recombinants with hemagglutinin (HA)-tagged PIF4, -6, and -7 were generated using bacmid technology (Fig. 2A), and anti-HA antibody was used instead.

FIG 1.

Identification of the components of the baculovirus PIF complex. (A to C) Identification of PIF complexes in AcMNPV (A), BmNPV (B), and HearNPV (C). The envelope fractions of wt AcMNPV, BmNPV, and HearNPV ODVs were subjected to BN-PAGE and analyzed by Western blotting with the indicated antibodies against PIFs. The ∼500-kDa PIF complexes are indicated by arrows. PIF8 dimers are indicated by closed arrowheads and PIF5 by open arrowheads. (D) Summary of the LC-MS/MS data on the PIF complexes. The numbers of identified proteins are shown. (E) Verification of other proteins in the AcMNPV PIF complex. BN-PAGE and Western analysis were conducted using the indicated antibodies, and anti-AcPIF1 was used as a positive control for identification of the PIF complex. Except for Ac108, none of the other proteins were identified as components of the PIF complex. ODV envelope proteins derived from appropriate OB concentrations (2 × 10⁷ OBs for detection of PIF5 and PIF8 and 1 × 10⁸ OBs for the other PIFs) were loaded into each lane for Western blot analysis. For staining, proteins from 3 × 10⁹ OBs were loaded.

FIG 2.

Construction of recombinant AcMNPVs. (A) Schematic diagrams for construction of recombinant AcMNPVs with HA-tagged PIF4, -6, and -7 by use of the Bac-to-Bac system. (B to K) Strategies for knocking out 10 pif genes. The A fragments containing pif genes and their modification strategies are shown. Transcriptional start sites and orientations of genes are indicated by long arrows; translational start codons (ATG or complementary CAT) are shown in green, the mutated or inserted nucleotides are shown in red, and translation was abolished by incorporating a stop codon as indicated.

BN-PAGE and Western blot analysis confirmed that the AcMNPV PIF complex was approximately 500 kDa (Fig. 1A), as reported previously (3, 22, 23). By use of anti-AcMNPV PIF antibodies, a similar, ∼500-kDa complex was detected in the closely related virus BmNPV (Fig. 1B). Moreover, antiserum raised against HearNPV PIF proteins confirmed the presence of a similar complex in the virus (Fig. 1C). These data suggest that an ∼500-kDa PIF complex in the envelope of ODVs is likely to be conserved in baculoviruses.

Consistent with previous reports (3, 22, 23), the AcMNPV PIF complex contained eight PIFs (PIF0 to -4 and PIF6 to -8) and lacked PIF5 (Fig. 1A). In addition to being in the PIF complex, PIF8 is also present in the ODV envelope, as an approximately 220-kDa protein (Fig. 1A). The predicted size of PIF8 is ∼96 kDa, but it was detected as a 110-kDa protein via SDS-PAGE (13); we speculate that the ∼220-kDa protein may be a dimer of PIF8. It is worth noting that in the envelopes of BmNPV and HearNPV, PIF5 was also absent from the complexes, and a putative dimer of PIF8 was detected (Fig. 1B and C).

To identify other possible components, proteomic analyses were carried out on the ∼500-kDa PIF complexes from AcMNPV, BmNPV, and HearNPV. LC-MS/MS analyses identified 25 viral proteins shared by all three viruses, including the 9 PIFs and 16 other proteins (Fig. 1D; Table 1). Among these proteins, ubiquitin, P24, and Ac26 were detected in insignificant amounts (Table 1) and were previously deemed to not be involved in oral infection (2), and they were therefore excluded from further analysis. Antibodies against the other 13 proteins were raised, but only anti-Ac108 recognized the PIF complex (Fig. 1E), and anti-PIF1 was used as a positive control in the experiment. Ac108, an 11.8-kDa protein encoded by ac108, is a homolog of Spodoptera frugiperda MNPV (SfMNPV) Sf58 (26) and BmNPV Bm91 (27). By using anti-Ac108 antiserum, we detected the presence of its homolog in the BmNPV PIF complex (Fig. 1B). An ac108-knockout mutant (Fig. 2K) had no effect on infectious BV production (Fig. 3A) or OB morphogenesis (Fig. 3B), but the knockout virus lost its oral infectivity to Spodoptera exigua larvae (Table 2). Therefore, ac108 was proven to encode an authentic PIF, which was designated PIF9. None of the other 12 remaining viral proteins were detected in the AcMNPV PIF complex (Fig. 1E), even though they were detected in the ODV envelope by SDS-PAGE and Western blot analysis (data not shown). This suggests that there are no other viral proteins in the complex. In summary, the ∼500-kDa PIF complex of AcMNPV consisted of nine PIF proteins: PIF0 to -4 and PIF6 to -9. As the accumulated size of the nine PIFs (∼360 kDa) is smaller than the observed size of 500 kDa, we further studied the contribution of each individual component to the formation of the entire complex.

TABLE 1.

Viral proteins identified in the PIF complex band of AcMNPV, BmNPV, or HearNPV by LC-MS/MS^a

Protein	AcMNPV				BmNPV				HearNPV
	Test 1		Test 2		Test 1		Test 2		Test 1		Test 2
	Coverage (%)	No. of unique peptides	Coverage (%)	No. of unique peptides	Coverage (%)	No. of unique peptides	Coverage (%)	No. of unique peptides	Coverage (%)	No. of unique peptides	Coverage (%)	No. of unique peptides
PIF0	47.75	30	47.91	30	55.81	37	55.81	37	38.08	24	40.26	29
PIF1	64.91	27	57.92	24	63.57	31	63.95	31	42.99	18	57.20	20
PIF2	56.02	23	56.02	23	55.26	21	66.72	23	59.53	20	56.02	23
PIF3	48.04	7	48.04	7	50.98	8	50.95	8	31.16	4	31.16	7
PIF4	9.25	1	25.43	2	21.43	3	21.43	3	28.90	6	28.90	6
PIF5	55.05	13	44.62	12	46.13	15	45.33	15	27.4	11	49.44	11
PIF6	36.46	7	34.38	5	55.22	7	55.22	7	54.89	8	58.65	7
PIF7	37.50	1	37.50	1	30.51	1	30.51	1	36.21	2	37.93	3
PIF8	63.99	39	65.76	42	52.21	34	53.87	32	53.55	34	60.17	38
Ac108 (PIF9)	49.52	4	49.52	4	49.52	4	49.52	4	53.19	4	51.06	4
ODV-E66	54.69	23	59.8	22	16.10	8	5.13	2	48.07	28	63.84	32
ODV-EC43	77.18	28	77.18	26	70.59	28	69.57	25	56.51	17	56.51	19
49K	69.60	35	74.18	34	68.27	37	66.39	31	61.97	31	57.48	27
ODV-E25	64.47	14	55.70	9	49.12	10	49.12	10	47.39	12	55.22	10
GP41	69.93	23	69.93	21	57.32	21	58.56	20	64.6	14	50.93	17
ODV-E18	16.13	1	16.13	1	33.66	2	33.66	2	45.68	3	51.85	3
VP80	60.93	45	59.77	45	57.95	50	58.81	48	29.09	16	20.50	12
Ac66	33.79	23	38.99	26	35.03	29	38.14	28	58.09	43	57.32	44
P33	32.82	10	40.15	10	27.03	8	22.78	8	42.52	13	49.21	18
Polyhedrin	37.96	9	43.67	11	46.12	10	57.14	12	48.78	13	46.75	14
VP39	19.31	5	25.65	5	10	3	10.29	3	12.29	3	15.02	9
Ac81	40.34	11	40.35	11	31.62	8	31.62	8	36.93	8	26.56	8
Viral ubiquitin	28.57	2	7.79	1	68.83	5	49.35	3	18.07	2	18.07	1
P24	50.51	9	57.58	9	14.87	3	57.58	9	14.92	3	3.63	2
Ac26	21.71	3	21.71	3	15.50	2	9.3	1	15.04	2	—	—
Ac73	69.70	5	43.35	11	18.18	2	18.18	2	—	—	—	—
Ac114	55.90	21	49.76	21	53.54	19	49.06	18	—	—	—	—
Ac78	33.03	3	33.03	3	—	—	—	—	17.27	2	22.73	4
P45	29.72	11	26.36	11	29.46	11	26.15	5	—	—	—	—
Ac74	36.23	8	34.34	8	—	—	—	—	—	—	—	—
Protein tyrosine phosphatase	41.67	5	50.00	8	15.48	3	—	—	—	—	—	—
35K	11.37	4	13.04	4	15.72	5	11.71	4	—	—	—	—
Ac5	44.04	4	67.89	6	45.87	4	45.87	4	—	—	—	—
Ac82	8.89	2	4.44	1	—	—	—	—	—	—	—	—
ssDNA binding protein	9.81	3	14.24	4	—	—	—	—	—	—	—	—
Ac145	36.36	2	27.27	1	22.10	1	26.32	1	—	—	—	—
Ac60	16.09	2	8.05	1	—	—	—	—	—	—	—	—
GP16	14.15	2	7.55	1	—	—	—	—	—	—	—	—
GP64	5.86	3	13.87	5	—	—	—	—	—	—	—	—
LEF-3	7.53	4	10.13	4	—	—	—	—	—	—	—	—
Ac91	—	—	—	—	10.39	1	10.39	1	—	—	—	—
Ac51	—	—	—	—	8.78	3	—	—	31.44	5	27.84	4
ME53	—	—	—	—	3.10	2	2.12	1	—	—	—	—
Ac75	—	—	—	—	7.52	1	7.52	1	—	—	—	—
LEF-12	—	—	—	—	—	—	—	—	20.18	4	—	—
Ha100	—	—	—	—	—	—	—	—	33.33	16	24.31	13
Ha44	—	—	—	—	—	—	—	—	15.08	6	14.29	8
Ha83	—	—	—	—	—	—	—	—	21.21	2	23.64	6
PP78/83	—	—	—	—	—	—	—	—	14.53	4	2.91	1

^a—, the protein was not identified in the sample by LC-MS/MS.

FIG 3.

Deletion of ac108 (pif9) had no effects on infectious BV production and OB morphogenesis. (A) One-step growth curves for Ac-WIV-Syn1 and AcΔpif9. Sf9 cells were infected with Ac-WIV-Syn1 and AcΔpif9 at an MOI of 5 TCID₅₀ units/cell, cell supernatants were harvested at the indicated time points, and BV titers were measured by endpoint dilution assay. Data points are mean values of titers derived from three independent infections. Error bars represent the standard deviations. (B) Electron microscopy of OBs of Ac-WIV-Syn1 and AcΔpif9. Upper and lower panels represent the results of SEM and TEM, respectively.

TABLE 2.

Feeding assay of OBs (3 × 10⁹ OBs ml⁻¹) on 4th-instar S. exigua larvae

Virus	Mortality (no. of dead larvae/total no. of larvae)
	Test 1	Test 2
Mock	0/24	0/24
Ac-wt	24/24	24/24
AcΔpif0	0/24	1/24
AcΔpif1	0/24	0/23
AcΔpif2	1/24	0/23
AcΔpif3	0/24	0/24
AcΔpif4	0/23	0/21
AcΔpif5	0/23	0/21
AcΔpif6	0/23	0/24
AcΔpif7	2/24	1/24
AcΔpif8	0/23	0/24
AcΔpif9	1/20	0/22

Contributions of individual PIFs to complex formation.

Previous reports showed that deletion of certain PIFs led to the loss of an intact PIF complex (13, 20, 22, 23, 28). Because those results were interpreted from a partially denatured complex by SDS-PAGE, which has a relatively low resolution, we decided to systematically explore the contributions of all nine PIFs to complex formation by BN-PAGE and Western blot analyses. We used deletion or point mutation strategies to inactivate pif genes from the AcMNPV genome (Fig. 2B to K); the resulting recombinants are all referred to as knockout mutants in this report. Individual pif-knockout AcMNPV mutants were generated by synthetic baculovirus technology (29) (see Fig. 4A for an example of a synthesis flowchart [for AcΔpif1]; details of the knockout strategy are shown in Fig. 2B to K). The resulting viruses were authenticated by sequencing and the absence of an immune response to PIF-specific antibodies by SDS-PAGE (Fig. 4B). None of the deletion mutants affected BV production (data not shown), but as expected, they all lost oral infectivity (Table 2).

FIG 4.

Generation and verification of pif-knockout AcMNPVs. (A) Schematic illustrating the generation of pif1-knockout AcMNPV. Fragment A36, which contains pif1, was modified by overlapping PCR to knock out pif1. After three rounds of transformation-associated recombination (TAR) in yeast cells, the synthetic genome of AcΔpif1 was obtained. The detailed strategies for knocking out the pif genes are shown in Fig. 2B to K. (B) Western blots of ODVs for identification of synthesized pif-knockout AcMNPVs. The envelope protein ODV-E25 and the nucleocapsid protein VP39 were used as protein loading controls. In order to inactivate endogenous protease in larva-derived OBs, they were heated at 80°C for 30 min prior to ODV purification.

We then analyzed the impacts of these pif-knockout mutants on PIF complex formation (Fig. 5). As expected, knockout of pif5 (Fig. 5A) had no effect on the integrity of the ∼500-kDa PIF complex. The PIF8 dimer was also detected and was associated with PIF9. On the other hand, pif8 knockout (Fig. 5B) led to a smaller complex, of approximately 400 kDa, containing at least PIF1 to -3 and PIF9 (PIF4, -6, and -7 were not tested due to a lack of antibodies). When either pif1, pif2, or pif3 was deleted, no complex was detected, except for the PIF8 dimer and its associated PIF9 (Fig. 5C to E). Knockout of pif0, pif4, pif6, pif7, or pif9 (Fig. 5F to J) resulted in an ∼230-kDa complex consisting of PIF1, -2, and -3. The PIF8 dimer was detected in these mutants, and PIF9 was found to be associated with the PIF8 dimer in the case of pif0, pif4, and pif6 deletions (Fig. 5F to H). The PIF8 dimer was slightly smaller than the ∼230-kDa complex, and this was more obvious when phosphate-buffered saline (PBS) was used, as represented by Fig. 5K. We assume that under PBS conditions the conformation of the complexes may be more natural and therefore results in better resolution by BN-PAGE. The picture that emerges from these experiments indicates a core complex of ∼230 kDa, consisting of PIF1 to -3, and a subcomplex of about 400 kDa when pif8 is deleted.

FIG 5.

Contributions of individual PIFs to PIF complex formation. (A to J) ODVs of AcΔpif0 to AcΔpif9 were purified in the presence of SBTI, and the ODV pellets were resuspended with ddH₂O and then subjected to envelope protein extraction. BN-PAGE and Western blotting were performed on the AcΔpif0 to AcΔpif9 ODV envelope fractions, using the indicated antibodies for probing. (K) BN-PAGE and Western blotting were conducted under the same conditions as those for panels A to J, except that the purified ODVs were resuspended in PBS (instead of ddH₂O). The PIF8 dimer was found to be smaller than the core complex. AcΔpif0 is shown as an example here, but similar results were obtained with AcΔpif4, AcΔpif6, AcΔpif7, and AcΔpif9.

Assembly of the PIF complex in the cytoplasm.

When the PIF complex was compromised by deletion of one of the PIFs, the other PIFs were still found in the ODV envelope (Fig. 4B). The question is whether the PIF complex is assembled in the nucleus or initially in the cytoplasm. To address this issue, cells were infected with AcMNPV and subjected to nuclear and cytoplasmic fractionation. Detection of the cytoplasmic protein marker glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (absent in the nucleus) and the nuclear marker lamin B (absent in the cytoplasm) showed that the fractionation was successful (Fig. 6A). The major BV envelope fusion protein GP64 (cytoplasmic) and two viral nucleocapsid proteins, VP39 and VP80 (nucleocytoplasmic), were also detected, further confirming the successful virus infection and cell fractionation (Fig. 6A). The cytoplasmic fractions were subjected to BN-PAGE and Western analysis (Fig. 6B). Anti-PIF1, -PIF2, -PIF3, -PIF8, and -PIF9 antibodies indicated that the ∼500-kDa complex was formed in the cytoplasm (Fig. 6B). It is interesting that the PIF8 dimer was already associated with PIF9 in the cytoplasm. Next, a coimmunoprecipitation (Co-IP) experiment was further conducted with anti-PIF1 to confirm the existence of a PIF complex in the cytoplasm. The results showed that the components of the PIF complex (PIF0, PIF1, PIF2, and PIF8), but not unrelated ODV envelope proteins (PIF5 and ODV-E66), were immunoprecipitated by anti-PIF1 antibody as confirmed by Western blot analysis (Fig. 6C). These results showed that the PIF complex is formed in the cytoplasm, possibly prior to transport into the nucleus.

FIG 6.

The PIF complex is assembled in the cytoplasm of infected cells. (A) Cytoplasmic and nuclear fractionation. Sf9 cells were infected with wild-type AcMNPV at an MOI of 5 TCID₅₀ units/cell, and cells were harvested at 72 h p.i. and separated into nuclear and cytoplasmic fractions. The purification of the fractions was analyzed by SDS-PAGE followed by Western blotting with the indicated antibodies. Lamin B and GAPDH were used as a cellular nucleus marker and a cytoplasmic marker, respectively. (B and C) PIF complexes could be detected in the cytoplasmic fraction. The cytoplasmic fractions from panel A were subjected to blue native PAGE (B) or used for Co-IP assay with mouse PIF1 antibody (C) and probed with the indicated antibodies. The complex in the BN-PAGE gel is indicated with an arrow.

The full complex, but not the subcomplex, is resistant to proteolytic degradation.

The alkaline environment in the insect midgut induces the dissolution of OBs to release occluded ODVs. Since proteolytic enzymes are ubiquitous in the insect midgut (30), ODVs need to resist proteolytic digestion prior to entry into epithelial columnar cells. It was reported previously that the PIF complex is important in protecting PIF proteins from degradation by OB endogenous proteinase (23). We further investigated the resistance of PIF proteins to proteolysis in the different forms of the complex.

Individual pif deletions (except for pif5 deletion) led to the rapid degradation of other PIF proteins during the process of ODV purification (Fig. 7A). In fact, we had to add proteinase inhibitors to obtain sufficient protein amounts for experiments with PIF deletion mutants (Fig. 5). The results suggest that only the full complex provides sufficient protection from proteolytic degradation for PIF proteins in the larval midgut during oral infection.

FIG 7.

The full complex, but not the subcomplex, confers PIF stability. (A) OBs of Ac-wt or pif-deletion AcMNPVs were left untreated prior to ODV purification. Western blots show the remaining PIFs in ODVs. AcΔpif8 was not included in the figure due to the presence of fewer ODVs, but it showed degradation patterns similar to those of the other pif deletion mutants. (B) Western blots of Ac-wt and AcΔpif7 ODVs purified in the presence of SBTI and exposed to proteinase K (PK; 40 ng/ml) or gut lumen fluid (0.2% [vol/vol]). Full-sized PIF0 and PIF8 are indicated with solid arrows, while the cleaved proteins are indicated with dashed arrows. (C) The intact, ∼500-kDa PIF complex (indicated by an arrow) of Ac-wt was still detected by use of PIF1 antibodies during BN-PAGE analysis after digestion as described for panel B, while the subcomplex of AcΔpif7 was barely detectable after proteinase K or gut lumen fluid treatment.

To further verify this hypothesis, ODV envelope proteins of wild-type AcMNPV (Ac-wt) and AcΔpif7 (as an example of pif knockouts) were purified in the presence of soybean trypsin inhibitor (SBTI) to avoid degradation by OB endogenous proteases. The envelope fractions of ODVs were extracted and exposed to proteinase K or S. exigua gut fluid. In the presence of proteinase K (40 ng/ml) or gut fluid (0.2% [vol/vol]), PIF proteins from Ac-wt were more resistant to proteolysis than those from AcΔpif7 (Fig. 7B). BN-PAGE and Western blot analyses showed that the intact, ∼500-kDa PIF complex of Ac-wt could still be detected with anti-PIF1 antibody under different treatment conditions; however, in the case of AcΔpif7, the subcomplex of ∼230 kDa was barely detectable after digestion (Fig. 7C). These results suggest that the full PIF complex is resistant to proteolytic digestion and remains intact prior to attachment to and entry into columnar epithelial cells.

DISCUSSION

Knowledge of the composition and structure of the PIF complex is fundamental to understanding its mode of action. Since PIFs were previously identified individually, we decided to take an integrated investigation to disclose all the possible components of the PIF complex. To this end, three different baculoviruses were chosen. AcMNPV and BmNPV both belong to the group I alphabaculoviruses, while HearNPV is a member of the group II alphabaculoviruses. Our results showed that PIF complexes of all three baculovirus are ∼500 kDa (Fig. 1), indicating that the size and composition of the PIF complex are conserved in members of the genus Alphabaculovirus, and possibly in all baculoviruses. LC-MS/MS analysis of these three PIF complexes revealed 25 shared proteins as potential candidates for complex components (Table 1). Western blot analysis confirmed the previous finding that PIF0 to -4 and PIF6 to -8 are constituents of the complex, and Ac108 was revealed as a novel component.

Ac108 is a homolog of SfMNPV Sf58, whose deletion abolished oral infectivity of SfMNPV in S. frugiperda (26). Previously, Ac108 was found to be associated with PIF1 by LC-MS/MS (22). Our results showed that it exists in the PIF complexes of both AcMNPV and BmNPV (Fig. 1B and E). Its deletion had no effects on BV production and OB morphogenesis (Fig. 3) but resulted in disassembly of the full PIF complex (Fig. 5J) and the loss of oral infectivity of AcMNPV in S. exigua (Table 2). Therefore, it is a genuine PIF and is designated PIF9. It is important to note that the previously identified nine PIFs (PIF0 to -8) are conserved in all sequenced baculoviruses, to date. PIF9, however, is conserved only in alpha-, beta-, and gammabaculoviruses and may thus be considered an auxiliary PIF in the complex and essential in some cases. The PIF9 homolog of BmNPV, Bm91, is dispensable for oral infectivity (27), which also alludes to an auxiliary function of this PIF in BmNPV ODV entry.

Because other viral proteins identified by LC-MS/MS were not found to be associated with the ∼500-kDa complex (Fig. 1E), it is convincing that the nine PIFs (PIF0 to -4 and PIF6 to -9) constitute all of the PIF components of the PIF complex. The documentation of the full members of the PIF complex is important for further investigations of its structure, stoichiometry, and assembly.

We next analyzed the impacts of individual PIFs on complex formation to gain insight into the structure of the complex. Notably, deletion of any pif gene other than pif5 eliminated the production of the full 500-kDa complex (Fig. 5). In detail, deletion of pif8 yielded an ∼400-kDa complex, suggesting that PIF8 is present as a monomer (∼110 kDa) in the complex and that its absence does not affect the assembly of other components. Deletion of pif0, -4, -6, -7, or -9 reduced the complex to ∼230 kDa. Deletion of pif1, -2, or -3 abrogated the assembly of the complex. Taken together, the data suggest a stable core complex of ∼230 kDa containing PIF1, -2, and -3, which appears to be a prerequisite for full complex assembly. We postulate that during assembly the core complex may form initially and serve as a docking station for other constituents to either form a full complex directly (Fig. 8A, left panel) or assemble the ∼400-kDa intermediate complex at first and subsequently add PIF8 to form the ∼500-kDa complete complex (Fig. 8A, right panel).

FIG 8.

Proposed model of PIF complex formation and intracellular transport. (A) Process of PIF complex formation. First, PIF1, -2, and -3 form a core complex which serves as a docking station, and then other constituents are recruited to the core, either directly to form a full complex (left) or via an about 400-kDa intermediate complex to which PIF8 is added to form the full PIF complex (right), which consists of nine PIFs. (B) Intracellular formation and transportation of the PIF complex. There are two possibilities for PIF complex formation and transportation, as follows. (1) The PIF complex is formed in the ER, and the INM-SM of the PIFs directs the PIF complex into the inner nuclear membrane; or (2) individual PIFs are transported into the nucleus via INM-SM and assembled in the nucleus. Either way, the entire PIF complex ends up at microvesicles (MVs) and is finally incorporated into the envelopes of ODVs. The ODVs are then embedded into OBs. VS, virogenic stroma.

Our result is different from previous postulates of the AcMNPV core complex, which was thought to be ∼170 kDa and to contain PIF1 to -4 (3, 22). It is now clear that PIF4 is not a component of the core complex, as deletion of pif4 still yields an ∼230-kDa complex (Fig. 5G). The difference between current results and previous data can be explained as follows. (i) The relatively low resolution of partially denaturing SDS-PAGE made it difficult to distinguish large proteins; the ∼150-kDa band, for example, may represent the ∼230-kDa core complex. (ii) As mentioned previously, the partial denaturation method could not entirely inhibit proteolytic degradation by OB endogenous proteinase (23). In our experience, different batches of larvae may have different activities of endogenous proteinases. These proteinases would make the subcomplex unstable if OBs were not pretreated with heat or proteinase inhibitors. (iii) The difference in pretreatments of the samples (SBTI inhibition or heating) may also have different impacts on the forms of the complex.

The nine PIF proteins in the complex have an accumulated size of ∼360 kDa, less than the observed ∼500 kDa. We cannot unequivocally exclude the possibility that there are one or more host proteins associated with the complex. However, it is more likely that some PIF proteins are present as multimers. The ∼230-kDa core complex is larger than the estimated size (∼127 kDa) of individual PIF1, -2, and -3 subunits, and we therefore postulate that some may be oligomers. Assuming that the ∼230-kDa core complex consists of PIF1, -2, and -3, the presence of PIF8 (∼110 kDa) and other monomer-form PIF components (total, ∼130 kDa) adds up to about 470 kDa, which is close to the observed ∼500-kDa size of the PIF complex. Due to the lack of specific antibodies against PIF4, -6, and 7, we are as of yet unable to propose a convincing composition model of the complex based on the current data. Future experiments using techniques such as crystal structure analysis and/or cryo-electron microscopy will be needed to gain a more detailed structure of the PIF complex.

ODV nucleocapsids are assembled and enveloped in the nuclei of infected cells, and given the size of the PIF complex, we were a bit surprised to find that the full complex already exists in the cytoplasm (Fig. 7). Taking the stoichiometry of the complex and the amount produced in the cytoplasm, the PIF complex found in the cytoplasm likely reflects the genuine process of PIF complex assembly, although we cannot exclude the possibility that it might be formed incidentally during the experimental process. It is now conceivable that the PIF proteins are transported as a complex from the cytoplasm to the nucleus. Many PIFs are rich in conserved cysteines (4), implying the requirement of protein folding machinery for correct assembly. Formation of the PIF complex in the cytoplasm may take advantage of such cellular machinery in the endoplasmic reticulum (ER). The INM-SM of the PIFs may direct the PIF complex into the inner nuclear membrane, intranuclear microvesicles (MVs), and eventually the envelopes of ODVs as INM-SM-containing proteins (31). Our results, however, do not exclude the possibility that individual PIFs can be transported into the nucleus and are assembled into complexes there. It is notable that the monomeric PIFs, when not in a complex due to pif knockout, apparently also move to the nucleus, as evidenced on the envelopes of ODVs (Fig. 4B). This is likely due to the presence of INM-SM in the PIFs (32). In the future, it will be interesting to reveal the mechanism for how exactly the PIF complex is assembled and transported to the nucleus. Based on current knowledge, a model of PIF complex formation and intracellular transport is proposed and illustrated in Fig. 8.

Our results confirmed the previous finding that the PIF complex protects PIF proteins from proteolytic degradation in the larval midgut (23) (Fig. 7). In addition, we showed that deletion of any component of the PIF complex resulted in disruption of the full complex structure (Fig. 5) and in protein degradation (Fig. 7A). This demonstrates that each individual component is required for complex formation to allow successful oral infection. Apart from protecting proteins from degradation, the essential role of the PIF complex is expected to include binding and fusion activities. So far, it is unknown which PIF is responsible for binding and/or fusion activities. For future investigations, we need to distinguish the role of each individual PIF from its impact on the integrity of the PIF complex. Since our results highlighted the important role of the core complex in the assembly of the full complex, our next goal is to understand how the core complex is formed and how it interacts with other PIFs to form the full complex.

MATERIALS AND METHODS

Amplification and purification of viruses.

Ac-wt and AcMNPV-WIV1-Syn1 (29) were propagated by orally infecting fourth-instar S. exigua larvae. AcMNPVs with pif deletions were amplified by intrahemocoelic injection of BV into S. exigua larvae. HearNPV (33) was amplified in fourth-instar H. armigera larvae, and BmNPV (34) was amplified in fourth-instar B. mori larvae. OBs were collected from diseased larvae and purified as previously described (35). ODVs were purified from OBs as previously reported (24), with minor modifications. To prevent protein degradation, SBTI (Sigma-Aldrich, St. Louis, MO) at a final concentration of 1 mg/10⁹ OBs was added prior to treatment with alkaline DAS buffer (0.1 M Na₂CO₃, 0.15 M NaCl, and 10 mM EDTA; pH 11.0). After DAS treatment at room temperature for 5 min, the solution was neutralized with a 1/10 volume of 500 mM Tris-HCl (pH 7.5). The ODV suspensions were subjected to sucrose gradient (30 to 60% [wt/vol]) ultracentrifugation at 72,000 × g for 1 h (SW28 rotor; Beckman Coulter, Brea, CA), after which ODV-containing bands were collected and diluted with double-distilled water (ddH₂O), followed by centrifugation at 72,000 × g for 1 h. The ODV pellets were then resuspended in ddH₂O for subsequent experiments. ODV purification in this study was performed as described above unless otherwise indicated.

Antibodies.

The rabbit polyclonal antibodies (pAbs) anti-AcPIF0, anti-AcPIF1, anti-AcPIF2, anti-AcPIF5, anti-AcODV-E66, anti-AcODV-E25, anti-AcGP41, anti-GAPDH, anti-lamin B, anti-AcGP64, anti-AcVP80, anti-AcP33, anti-Ac-polyhedrin, anti-AcVP39, anti-HaPIF0, anti-HaPIF1, anti-HaPIF2, and anti-HaPIF3 were produced in our lab as reported previously (36–39). Other antibodies are reported here for the first time; these include the rabbit pAbs anti-AcPIF8 (amino acids [aa] 466 to 847), anti-AcEC43, anti-Ac49K, anti-AcVP80, anti-Ac66, anti-Ac81 (aa 15 to 164), anti-HaPIF5 (aa 1 to 318), and anti-HaPIF8 (aa 180 to 816) as well as the mouse pAbs anti-AcPIF1 (aa 24 to 530) and anti-AcODV-E18 (aa 20 to 62). These pAbs were generated by injecting rabbits or mice with purified prokaryotically expressed full-length or truncated proteins (as indicated above). The monoclonal antibodies (MAbs) anti-AcPIF3 (33A6), anti-AcPIF9 (26A9), and anti-HaPIF6 (38C10) were isolated after 3 rounds of screening of mouse hybridoma cells. An HA tag MAb and horseradish peroxidase-conjugated secondary antibodies against rabbit or mouse IgG were purchased from Sigma-Aldrich.

BN-PAGE and LC-MS/MS analysis.

ODV suspensions were incubated with an equal volume of TN buffer (100 mM Tris, pH 7.4, and 2% NP-40) in the presence of a proteinase inhibitor cocktail (Roche, Basel, Switzerland) at 4°C for 30 min. After centrifugation at 20,000 × g for 30 min, the supernatant was collected in a new tube and mixed with 4× NativePAGE sample buffer (200 mM Bis-Tris, 64 mM HCl, 200 mM NaCl, 40% [wt/vol] glycerol, and 0.004% Ponceau S; pH 7.2), after which 5% Coomassie brilliant blue G-250 was added to the sample at a final concentration of 1/4 the concentration of NP-40. The samples were analyzed by BN-PAGE or stored at −80°C until further use. Electrophoresis and staining were performed with a NativePAGE Novex Bis-Tris gel system (Life Technologies, Carlsbad, CA) according to the manufacturer’s instructions. For Western blot analysis, proteins were semidry blotted (Bio-Rad Laboratories, Hercules, CA) with Tris-glycine buffer and developed by MicroChemi (DNR Bio-Imaging Systems, Neve Yamin, Israel).

Visible bands for the PIF complex were excised from the gel and treated with 25 mM ammonium bicarbonate in 50% (vol/vol) acetonitrile, followed by reduction, alkylation, and incubation with 0.01 mg/ml trypsin overnight. The digested peptides were extracted with 5% (vol/vol) formic acid in 50% (vol/vol) acetonitrile and vacuum dried. The digested samples were analyzed using Q Exactive spectrometers (Thermo Fisher Scientific, Waltham, MA) coupled with an Easy-nLC 1200 Nanoflow liquid chromatography system (Thermo Fisher Scientific). Tryptic digests were loaded onto a C₁₈ column (75-μm internal diameter and 25-cm length; Thermo Fisher Scientific) and eluted by use of a gradient of 5 to 35% (vol/vol) acetonitrile in 0.1% (vol/vol) formic acid at a flow rate of 300 nl/min for 90 min. The MS full-scan range was m/z 350 to 1,300, with resolutions of 700,000 at 200 m/z. The top 20 peaks were subjected to MS/MS analysis with resolutions of 17,500 at 200 m/z. The dynamic exclusion time was set to 18 s. Data were acquired using Xcalibur software (version 4.0; Thermo Fisher Scientific).

Proteins were identified by searching against NCBI databases by use of Mascot, version 2.3 (Matrix Science, London, United Kingdom), implemented in Proteome Discoverer 2.1 (Thermo Fisher Scientific). Accession numbers in the databases were NC_001623.1 (AcMNPV), NC_001962.1 (BmNPV), and NC_002654.2 (HearNPV). Proteins with at least two unique peptides with a 1% false-discovery rate at the peptide level were considered positive identifications.

Construction of recombinant AcMNPVs.

As no specific antibodies to AcMNPV PIF4, -6, or -7 were obtained, recombinant AcMNPVs with HA-tagged PIFs were generated by use of the Bac-to-Bac system as previously reported (9, 11, 12), with minor modifications (Fig. 2A). Briefly, the pif4-knockout bacmid was generated by inserting the chloramphenicol resistance gene (Cm^r) and an egfp cassette between nucleotides (nt) 386 and 389 of pif4 by homologous recombination in Escherichia coli. polyhedrin (polh) and pif4 fused with HA were used to repair the pif4-knockout bacmid by Tn7-mediated transposition at the polh locus. The bacmid was used to transfect Sf9 cells, and progeny viruses were harvested. After confirmation by PCR and sequencing, the correct virus was named Ac-pif4HA. As there is overlap between the sequences of lef3 and pif6, a double-knockout bacmid was first generated by replacing the sequence region covering nt 1 to 171 of pif6 and 1 to 792 of lef3 with the Cm^r and egfp cassettes to generate Ac-pif6HA. HA-tagged pif6 and lef3 together with polh were then used to repair the bacmid by transposition. For Ac-pif7HA, nt 29 to 145 of pif7 were replaced with the Cm^r and egfp cassettes.

To generate 10 individual pif-knockout viruses, a synthetic biology method similar to that for generating AcMNPV-WIV1-Syn1 (29) was employed (Fig. 2B to K and Fig. 4A). This method is based on three rounds of transformation-associated recombination (TAR) in yeast cells. Briefly, 45 overlapping fragments (A1 to A45; ∼3 kb each) covering the entire genome of AcMNPV were produced by PCR and recombined into nine fragments (B1 to B9; ∼15 kb each) by TAR in yeast. The B fragments were further recombined into three larger intermediates (C1 to C3; ∼45 kb each), and the complete AcMNPV-WIV1-Syn1 genome was finally obtained by TAR with three C fragments in yeast. The synthetic virus was rescued by transfecting Sf9 cells with genomic DNA. To generate pif0-knockout bacmids, the region of nt 119428 to 121229, located on fragments A41 and A42 of the AcMNPV genome (accession number KM667940), was deleted by overlapping PCR (Fig. 2B). The AcΔpif0 bacmid was obtained by three rounds of TAR as described above, but with modified A41 and A42. Similarly, for pif1 knockout, fragment A36 was modified by deleting nt 100709 to 102384 of AcMNPV to obtain the AcΔpif1 bacmid by three rounds of TAR in yeast (Fig. 2C). To generate AcΔpif2, AcMNPV nt 17388 to 18453, which contained the majority of the pif2 open reading frame (ORF), were deleted; however, part of the 5ʹ end of the gene, encoding the first 24 aa of pif2, was retained because it contains the transcriptional initiation motif of the neighboring gene arif1 (40) (Fig. 2D). To ensure that the remaining amino acids were not synthesized, the transcriptional initiation motif TAAG of pif2 was mutated to TGAG, and a G was inserted right after nt +4 of the translational start codon (ATG) to generate a frameshift mutation (Fig. 2D). To generate the pif3 knockout, nt 99281 to 99874 of AcMNPV were deleted in fragment A35 (Fig. 2E). To knock out pif4 without affecting neighboring genes, the transcriptional initiation motif TAAG of pif4 was mutated to GAGG. To further inhibit the synthesis of PIF4, a stop codon (TGA) was generated by mutating TTG just a few nucleotides downstream of the start codon, and an insertion of TA generated another stop codon (TAA) (Fig. 2F). To knock out pif5, AcMNPV nt 129091 to 130235 were removed from fragments A44 and A45 (Fig. 2G). In the case of pif6, the transcriptional initiation motif TAAG of pif6 was converted to TGAG, and a TGG just downstream of the ATG was mutated to the stop codon TGA. A second stop codon (TGA) was introduced when a C was converted to T. In addition, an inserted C generated a frameshift mutation (Fig. 2H). To knock out pif7, the transcriptional initiation motif TAAG was mutated to TCAG, and a T was inserted right after the translational start codon (ATG) to generate a stop codon (Fig. 2I). For pif8, the majority of the sequence was deleted, except for a fragment of 325 nt at the 5ʹ end and an essential region for nucleocapsid assembly in the middle of the gene (41). To ensure that the remaining sequence does not encode amino acids, the transcriptional initiation motif TAAG of pif8 was mutated to TAAA, and a T at nt 58 of the ORF was deleted to generate a stop codon and a frameshift mutation (Fig. 2J). Similar to that of pif7, pif9 knockouts were generated by mutating the transcriptional initiation motif TAAG to TCAG and adding a T right after the translational start codon (ATG) to generate a stop codon (Fig. 2K). After confirmation by sequencing, the synthesized bacmids were used to transfect Sf9 cells, and then progeny viruses were collected for further amplification and infectivity tests in vitro and in vivo. Viruses with individually inactivated pif genes were designated AcΔpif0 to AcΔpif9.

One-step growth curve assay.

Sf9 cells (3 × 10⁶) were infected with the viruses AcMNPV-WIV-Syn1 and AcΔpif9 at a multiplicity of infection (MOI) of 5 50% tissue culture infective dose (TCID₅₀) units/cell. At 0, 24, 48, 72, and 96 h p.i., the cell supernatants were collected and the virus titers were measured by endpoint dilution assay. Virus infections were performed in triplicate. BV titers at specified time points were analyzed statistically by one-way analysis of variance (ANOVA), and the results were presented using GraphPad Prism 6 software.

Electron microscopy.

Purified OBs of AcMNPV-WIV-Syn1 and AcΔpif9 were analyzed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) as previously reported (24). Briefly, OBs were dried on tinfoil paper and sputter coated with gold for SEM analysis (Hitachi SU-8010 instrument). For TEM analysis, OBs were fixed with 2.5% (vol/vol) glutaraldehyde at 4°C overnight and ultrathinly sectioned for examination by TEM (Hitachi H-7000FA instrument).

Feeding assays.

OBs of Ac-wt and AcΔpif0 to AcΔpif9 were purified from diseased larvae infected through BV injections and then used in feeding assays by the droplet method (42). Briefly, OBs were diluted with feeding buffer (1% [wt/vol] brilliant blue and 10% [wt/vol] sucrose) to a final concentration of 3 × 10⁹ OBs/ml and fed to 16-h-starved early-fourth-instar S. exigua larvae. Mortality was recorded at 3 days postinfection (p.i.). The feeding assay was performed twice independently.

Cytoplasmic and nuclear fractionation.

Fractionation of the cytoplasm and nucleus was performed according to the method in a previous report (11), with modifications. Sf9 cells (1.2 × 10⁸) were infected with wild-type AcMNPV budded virus at an MOI of 5 TCID₅₀ units/cell. At 72 h p.i., infected cells were harvested and washed with cold PBS three times. Cells were resuspended in 2 ml of NP-40 lysis buffer (PBS, 0.5% NP-40, proteinase inhibitor cocktail) and incubated on ice for 30 min. Lysis of the plasma membrane but not the nuclear membrane was confirmed by microscopy. Cell lysates were spun at 1,000 × g for 3 min. The supernatant (1.8 ml was taken for subsequent experiments, and the rest [0.2 ml] was discarded) and the pellet (washed with 1 ml lysis buffer and then resuspended with 2 ml lysis buffer) were collected as the cytoplasmic and nuclear fractions, respectively. GAPDH and lamin B were used as marker proteins to show the efficiency of fractionation. For BN-PAGE analysis of the cytoplasmic samples, addition of 4× NativePAGE sample buffer and digitonin (final concentration, 1%) plus G-250 and 10 s of sonication were required before sample loading.

Co-IP analysis.

Protein A beads (NewZongKe Viral Disease Control Bio-Tech LTD, China) were precoated with PIF1 mouse polyclonal antibody or preimmune mouse serum as a control and incubated with the AcMNPV-infected cytosolic fraction overnight at 4°C. After washing 4 times with lysis buffer, the precipitated proteins were analyzed by Western blotting using the different PIF antibodies.

Proteinase protection assays.

Gut lumen fluid was taken from the mouths of fifth-instar S. exigua larvae by squeezing their abdomens. Collected gut fluid was centrifuged at 2,400 × g for 5 min, and then the supernatant was harvested, divided into 10-μl aliquots, and stored at −80°C until further use. The proteinase activity of the collected gut fluid was determined by dose-dependent digestion of bovine serum albumin (BSA).

Envelope proteins were extracted from AcMNPV ODVs by incubating purified virions with 1% NP-40 in ddH₂O at 4°C for 30 min, followed by harvesting of supernatants after centrifugation at 20,000 × g for 30 min. The envelope fractions (ODV envelope proteins from 1 × 10⁹ OBs) were digested with proteinase K (40 ng/ml) or gut lumen fluid (0.2% [vol/vol]) at 37°C for 5 min. Proteins were loaded at 1 × 10⁸ OB-derived proteins per lane for Western blot analysis.