The F protein of Helicoverpa armigera single nucleopolyhedrovirus can be substituted functionally with its homologue from Spodoptera exigua multiple nucleopolyhedrovirus

The initial infection events of enveloped viruses are usually mediated by their envelope proteins. The envelope fusion proteins (EFPs) of baculoviruses play important roles in several steps in the infection processes, such as receptor binding, fusion with endosomal membrane and efficient budding (Hefferon et al., 1999; Kingsley et al., 1999; Oomens & Blissard, 1999). There are two types of EFP in baculoviruses: GP64 and F protein. GP64 is a glycoprotein encoded by group I nucleopolyhedroviruses (NPVs), such as Autographa californica multinucleocapsid NPV (AcMNPV) (Ayres et al., 1994), Orgyia pseudotsugata MNPV (OpMNPV) (Ahrens et al., 1997) and the Bombyx mori NPV (Gomi et al., 1999). F protein has been identified in group II NPVs such as Lymantria dispar MNPV (Pearson et al., 2000), Spodoptera exigua MNPV (SeMNPV) (IJkel et al., 2000) and Helicoverpa armigera singlenucleocapsid NPV (HearNPV) (Long et al., 2006a). F homologues have been found in all baculovirus genomes sequenced so far.

GP64 and F proteins differ in their structure and mode of action. In contrast to the high amino acid identity among GP64 proteins (>74 %), F proteins are more diverse (20–40 %) (Pearson et al., 2000). Phylogenetic analysis has suggested that F proteins may be an ancient type of EFPs (Pearson et al., 2000). In group I NPVs, there is a truncated form of F, which occurs in the budded virus (BV) phenotype, but is not responsible for fusion (Lung et al., 2003). It is essential for F proteins to be cleaved by cellular furin for activation, which generates a membrane-anchored F₁ subunit and membrane-distal F₂ subunit connected by a disulfide bond (IJkel et al., 2000; Westenberg et al., 2002). GP64 does not require furin cleavage for activation. Furthermore, disulfide bonds are not involved in the formation of F protein oligomers, as is the case for the GP64 trimeric structure (Long et al., 2006a; Oomens et al., 1995).

So far, several F proteins of group II NPVs have been demonstrated to be able to replace GP64 by pseudotyping AcMNPV (Long et al., 2006a; Lung et al., 2002). However, the reverse does not occur (Westenberg & Vlak, 2008). In this report, we addressed this issue by replacing the F protein of HearNPV (HaF) with its homologue from SeMNPV (SeF). HaF and SeF are the well-studied representatives of baculoviral F proteins (IJkel et al., 2000; Long et al., 2006a) and are typical class I fusion proteins. HaF and SeF use the same insect-cell receptor, whereas GP64 uses a different receptor (Westenberg et al., 2007). The amino acid identity of SeF and HaF is about 33 % but the two proteins have many common structural features such as an N-terminal signal peptide, furin cleavage site, fusion peptide region, transmembrane region (TM) and a cytoplasmic tail domain (CTD) (IJkel et al., 2000; Long et al., 2006b).

In this report, we describe the construction and testing of a HaF-knockout HearNPV bacmid, HaBacΔF. HaF and SeF were inserted into this HaF-null HearNPV bacmid, generating HaBacΔF-HaF (rescue control) and HaBacΔF-SeF, respectively. Transfection/infection assays and one-step growth curves were conducted to compare the BV production of HaBacΔF-HaF virus (vHaBacΔF-HaF) and vHaBacΔF-SeF. Low-pH-dependent membrane fusion assays, Western blot analysis and neutralization assays were performed to investigate the characteristics of the pseudotyped viruses and hence the function of F in a heterologous group II NPV.

Insect cells and virus.
HzAM1 cells were maintained at 27 °C in Grace's insect medium (Gibco-BRL) supplemented with 10 % fetal bovine serum (FBS) (pH 6.0). The HearNPV bacmid (HaBacHZ8) used in this study was constructed previously in our laboratory (Wang et al., 2003).

Construction of recombinant viruses.
The HaF gene of HaBacHZ8 was knocked out by homologous recombination in recA⁺ Escherichia coli BJ5183 cells according to the method of Hou et al. (2002), replacing the HaF gene by the tetracycline-resistance gene. Briefly, 1.1 kb of sequence upstream of the HaF gene was amplified by PCR with primers HaF-up-for (5'-GGTACCAGTTTCACAATTCATGTCGGGC-3'; KpnI site underlined) and HaF-up-rev (5'-AAGCTTCTGCAGACAAAACTGACGTTGAACAC-3'; HindIII site underlined). An 850 bp sequence downstream of the HaF gene was obtained with primers HaF-down-for (5'-GAGCTCGATTCTTTCGATCAATATGATG-3'; SacI site underlined) and HaF-down-rev (5'-GAATTCGTACAACCAATAGTATACCG-3'; EcoRI site underlined) using HearNPV as template. The PCR products were cloned into a pKS vector (Stratagene). An enhanced green fluorescent protein gene (egfp) under the control of the hsp70 promoter was further cloned into the pKS vector using the XhoI and HindIII sites. A tetracycline-resistance gene (Tet^R) was amplified using primers Tet-for (5'-TCATGTTTGACAGCTTATCATCGATAAGCTATAATGCGGTAGTTTATCACAGTT-3') and Tet-rev (5'-GGCTTCCATTCAGGTCGAG-3') using pBR322 (NEB) as template. The XhoI site of the pKS vector was blunted and Tet^R was cloned into this site, generating the transfer vector pKS-ΔF. This transfer vector was digested by KpnI and EcoRI and the linear fragment containing Tet^R, egfp and the HaF gene flanking sequences was used to transform BJ5183 competent cells containing HaBacHZ8 and helper plasmid pKD46 as described by Hou et al. (2002). Positive clones were selected by tetracycline and kanamycin resistance. The correct bacmid clone was authenticated by PCR and named HaBacΔF.

For the construction of pseudotyping viruses, the OpMNPV gp64 promoter (Op166) was digested from p166BRNX-AcV5 (IJkel et al., 2000) using SacI and EcoRI and cloned into pUC19 (TaKaRa) to give pUC19-Op166. The Op166 promoter was then excised by BamHI digestion and subcloned into the transfer vector pFastBac1 (Bac-to-Bac Baculovirus Expression System; Gibco-BRL), generating pFB-Op166. The HaF gene was amplified from the HearNPV DNA template by PCR with primers HaF-for (5'-AAGCTTATGGTTGCGATAAAAAGTAGTATG-3'; HindIII site underlined) and HaF-rev (5'-GGATCCAAGCTTCGTAGGGATTTGCCGTCG-3'; HindIII site underlined). The PCR product was digested with HindIII and cloned into the transfer vector pFB-Op166, generating pFB-Op166-HaF. The SeF gene was excised from p166AcV5-Se8 (IJkel et al., 2000) by BamHI and EcoRI digestion and cloned into BamHI/EcoRI-cleaved pFastBac1, generating pFB-SeF. The Op166 promoter was digested from pUC19-Op166 by BamHI and cloned into pFB-SeF, generating pFB-Op166-SeF. pFB-Op166-HaF or pFB-Op166-SeF was used to co-transform DH10Bac cells with a HaBacΔF bacmid and the helper plasmid expressing transposase (Bac-to-Bac Baculovirus Expression System; Gibco-BRL). Recombinant bacmids were selected by gentamicin resistance and blue/white screening. Correct recombinant bacmids were identified by PCR with M13 primers and EcoRI digestion, and named HaBacΔF-HaF and HaBacΔF-SeF, respectively.

Transfection and infection assays.
HzAM1 cells were inoculated into 35 mm diameter tissue culture dishes at a density of 5x10⁵ cells per dish. After 2 h, cells were transfected with 1 µg recombinant viral DNA (HaBacΔF, HaBacΔF-HaF or HaBacΔF-SeF) using 15 µl Lipofectin according to the Bac-to-Bac Expression Systems manual (Invitrogen). For the infection assay, at 6 days post-transfection (p.t.), 1 ml supernatant from the transfection was centrifuged at 956 g for 5 min to remove cell debris and the supernatants were used to infect HzAM1 cells. Cells were monitored by fluorescence microscopy at 72 h p.t. or post-infection (p.i.).

One-step virus growth curves.
HzAM1 cells were infected with vHaBacΔF-SeF or vHaBacΔF-HaF at an m.o.i. of 10 TCID₅₀ units per cell. At 0, 12, 24, 48 and 72 h p.i., supernatants were harvested and titrated by an end-point dilution assay. Each virus infection was done in triplicate. BV titres were log-transformed and statistically analysed with two-way analysis of variance in GLM (SPSS Inc., 2003) with virus type and time as factors.

Low-pH induced envelope fusion assay.
The syncytium-forming ability of the pseudotyped viruses was tested according to the method of Blissard & Wenz (1992) with a slight modification. Briefly, HzAM1 cells were infected by recombinant vHaBacΔF-HaF or vHaBacΔF-SeF at an m.o.i. of 10 TCID₅₀ units per cell. At 24 h p.i., the cells were washed twice with Grace's insect medium and then treated with acidic (pH 5.0) Grace's insect medium. After being exposed to the low-pH medium for 5 min, the cells were further cultured with normal Grace's insect medium containing 10 % FBS. Syncytium formation was observed under a fluorescence microscope at 24 h after the downward pH shift.

Western blot analysis of recombinant BVs and infected cells.
For Western blot analysis, anti-HaF₁ and anti-SeF₁ antibodies were used. To generate anti-HaF₁ antibody, the HaF₁ sequence fragment without the TM domain was amplified by PCR with primers HaF₁-for (5'-CAAGGATCCAAACATTGGATTGAACTTCGTTG-3'; BamHI site underlined) and HaF₁-rev (5'-AATAAGCTTATCCCGTACTTAAATTCCAACCGC-3'; HindIII site underlined). PCR products were cloned into a pET28a expression vector (Novagen). To generate anti-SeF₁ antibody, an SeF₁ sequence fragment without the TM region was amplified by PCR using forward primer 5'-GGGGGATTCATGGGCCTTTTTAATTTTATGGGAC-3' (BamHI site underlined) and reverse primer 5'-GGGAAGCTTTTACTTTACGTAATGAAAATCGATACC-3' (HindIII site underlined), and cloned into a pET28a expression vector. The expression plasmids containing HaF₁ and SeF₁ were transformed into BL21 cells and the proteins were induced with 1 mM IPTG at 37 °C for 3 h. The expressed HaF₁ and SeF₁ proteins were purified by continuous-elution electrophoresis using a Model 491 Prep Cell (Bio-Rad) and used to immunize rabbits to generate polyclonal antisera against these proteins.

Western blot analysis was performed to detect furin cleavage of F proteins in the recombinant BVs as well as in the infected cells. HzAM1 cells were infected with recombinant vHaBacΔF-HaF or vHaBacΔF-SeF at an m.o.i. of 10 TCID₅₀ units per cell, and SeUCR cells were infected by SeMNPV at the same m.o.i. At 3 days p.i., 2 ml supernatant containing fresh BVs was harvested and centrifuged at 956 g for 5 min to remove cell debris and then at 20 800 g for 30 min at 4 °C. The sedimented BVs and infected-cell samples were disrupted in 6x SDS-PAGE sample buffer and separated by 12 % SDS-PAGE. The proteins in the gel were transferred to Hybond-N membranes (Amersham) by semi-dry electrophoresis. Western blot analysis was performed with polyclonal anti-HaF₁ or anti-SeF₁ serum as primary antibody and alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin as secondary antibody; the signals were detected with NBT/BCIP (SABC).

Neutralization assays.
HzAM1 cells were inoculated into 24-well tissue culture plates at a density of 2x10⁵ cells per well. vHaBacΔF-HaF or vHaBacΔF-SeF (m.o.i. of 1 TCID₅₀ unit per cell) was incubated with different amounts of polyclonal rabbit anti-HaF₁ or anti-SeF₁ antiserum together with Grace's insect medium containing 10 % FBS to a final volume of 300 µl. For mock neutralization, only virus (vHaBacΔF-HaF or vHaBacΔF-SeF; m.o.i. of 1 TCID₅₀ unit per cell) was incubated with Grace's insect medium with 10 % FBS to a volume of 300 µl. The viruses were also incubated with pre-immune serum as controls. All of the mixtures were incubated for 1 h at room temperature and then added to monolayers in 24-well tissue culture plates for 1 h to allow virus attachment. The mixtures were then replaced with normal medium and infection rates were quantified at 60 h p.i. using a Beckman Coulter (EPICS XL) flow cytometer.

Generation of HaF-knockout HearNPV bacmid and pseudotyped bacmids
The HaF-knockout bacmid HaBacΔF and pseudotyped bacmids HaBacΔF-HaF and HaBacΔF-SeF were generated as described in Methods; their genomic organization is shown in Fig. 1(a). Positive clones were confirmed by PCR amplifications (data not shown) and EcoRI analysis (Fig. 1b). In HaBacHZ8, the HaF gene is located in an 8.9 kb EcoRI fragment. In HaBacΔF, the 8.9 kb band was changed to a 9.5 kb band due to substitution of the HaF gene with a Tet^R and egfp cassette (Fig. 1b). In comparison with HaBacΔF, the EcoRI profile of HaBacΔF-HaF contained additional 14.4 kb and 2.7 kb bands, but did not have the 12.2 kb band; these changes were due to insertion of the HaF gene (Fig. 1b). In HaBacΔF-SeF due to insertion of the SeF gene, a new band of 17.1 kb replaced the band of 12.2 kb in HaBacΔF (Fig. 1b). All of the bands of mutant bacmids were shifted as expected (indicated by arrowheads, Fig. 1b), confirming that the bacmid constructs were correct.

(40K): Fig. 1. Construction and identification of recombinant bacmids. (a) Schematic representation of the recombinant bacmids HaBacΔF, HaBacΔF-HaF and HaBacΔF-SeF, and their parent bacmid HaBacHz8. (b) Restriction enzyme analysis of the parental and recombinant bacmids. Recombinant bacmid DNAs were isolated and analysed by EcoRI digestion. Shifted bands are indicated by arrowheads (original bands with open arrowheads; shifted bands with filled arrowheads).

Transfection and infection assays
HzAM1 cells were transfected with recombinant bacmid HaBacΔF, HaBacΔF-HaF or HaBacΔF-SeF. At 72 h p.t., cells were observed by fluorescence microscopy (Fig. 2a–c). When cells were transfected with HaBacΔF, fluorescence was limited to single cells (Fig. 2a), whereas transfection with HaBacΔF-HaF or HaBacΔF-SeF resulted in primary as well as secondary infections (Fig. 2b, c). When transfection supernatants were used to infect a new batch of cells (Fig. 2d–e), only those from vHaBacΔF-HaF and vHaBacΔF-SeF contained infectious virus (Fig. 2e, f). The results thus showed that when HaF was eliminated from HearNPV, no infectious BV was generated (Fig. 2d), whilst both HaF and SeF rescued the infectivity of HearNPV (Fig. 2e, f). Therefore, HaF is an essential gene for secondary HearNPV infection, as is GP64 in the case of group I NPVs (Monsma et al., 1996). The presence of BV in the supernatant of HzAM1 cells transfected with vHaBacΔF-SeF could be detected by PCR analysis (data not shown) and by the infectivity for HzAM1 cells (Fig. 2e, f).

(59K): Fig. 2. Transfection/infection assay of recombinant HearNPVs. HzAM1 cells were transfected with bacmid HaBacΔF (a), HaBacΔF-HaF (b) or HaBacΔF-SeF (c). The supernatants of the transfection cells (a, b and c) were used to infect HzAM1 cells (d, e and f, respectively). Green fluorescence was detected at 72 h p.t. or p.i.

One-step growth curves of the recombinant viruses
To study the effect of the insertion of HaF and SeF in HaBacΔF on the growth characteristics of the pseudotyped viruses, one-step growth curve assays were conducted. HzAM1 cells were infected with vHaBacΔF-HaF or vHaBacΔF-SeF at an m.o.i. of 10 TCID₅₀ units per cell and growth curve assays were performed by end-point dilution. The data presented in Fig. 3 show that the two viruses had similar dynamics of BV production. However, at the late stages of infection, the BV titres of vHaBacΔF-HaF were higher than those of vHaBacΔF-SeF. For example, at 48 h p.i., the BV titre of vHaBacΔF-HaF was 7.43±0.98x10⁶ TCID₅₀ units ml^–1 and the BV titre of vHaBacΔF-SeF was 1.00±0.17x10⁶ TCID₅₀ units ml^–1, whilst at 72 h p.i., the BV titre of vHaBacΔF-HaF was 1.24±0.70x10⁷ TCID₅₀ units ml^–1 and the BV titre of vHaBacΔF-SeF was 1.70±0.52x10⁶ TCID₅₀ units ml^–1. Therefore, in the late phase of infection, the infectious BV production of vHaBacΔF-HaF was almost ten times that of vHaBacΔF-SeF. Statistical analysis showed that vHaBacΔF-SeF had a significantly decreased BV titre compared with the rescued virus vHaBacΔF-HaF (F = 92.375, d.f. = 1, 19, P < 0.01).

(17K): Fig. 3. One-step growth curves of the pseudotyped viruses. HzAm1 cells were infected with vHaBacΔF-HaF or vHaBacΔF-SeF at an m.o.i. of 10 TCID50 units per cell. BV titres at different time points were quantified by end-point dilution assays. Results are shown as means±SEM.

Syncytium formation assay of the pseudotyped viruses
As baculovirus F proteins are mildly acid-triggered membrane fusion proteins and cause low-pH-dependent membrane fusion, an assay was conducted to examine the expression and fusogenic ability of both F proteins (Fig. 4). As indicated by arrows in Fig. 4(c) and (d), multinuclear cells were detected in both vHaBacΔF-HaF- and vHaBacΔF-SeF-infected HzAM1 cells at 24 h after the pH was shifted down. Observation of the fusion level (number of syncytia) and the size of syncytia showed the fusion ability of vHaBacΔF-SeF to be similar to that of vHaBacΔF-HaF in infected HzAM1 cells.

(116K): Fig. 4. Syncytium formation assay. HzAM1 cells were infected with vHaBacΔF-HaF (a, c) or vHaBacΔF-SeF (b, d) at an m.o.i. of 10 TCID50 units per cell. At 48 h p.i., cells were incubated with Grace's insect medium at pH 6.0 (a, b) or 5.0 (c, d) for 5 min and then examined for syncytium formation after 24 h. Large syncytia containing many nuclei were identified and are indicated by arrows in (c) and (d).

Western blot analyses of the recombinant BVs and infected cells
To detect the F proteins in the recombinant BVs, Western blot analyses were conducted. Fig. 5(a) shows that the anti-HaF₁ antiserum detected the expected 59 kDa band for HaF (Long et al., 2006a) in vHaBacΔF-HaF BVs but not in vHaBacΔF-SeF BVs; the anti-SeF₁ antiserum detected the expected 60 kDa band for SeF (IJkel et al., 2000) in vHaBacΔF-SeF BVs but not in vHaBacΔF-HaF BVs. Antiserum against HearNPV nucleocapsid protein VP80 served as a control for equal amounts of BV protein on the blot. The results indicated that both HaF and SeF were correctly cleaved into subunits F₁ and F₂ incorporated into the BVs, as no F₀ bands were detected (Fig. 5a).

(70K): Fig. 5. Western blot analysis of F proteins in pseudotyped BVs and in infected cells. (a) Western blot analysis of recombinant BVs. BVs were harvested from supernatants of infected HzAM1 cells and Western blot analyses were performed using antibodies against HaF1, SeF1 and VP80. Lanes: M, protein molecular mass marker; 1, vHaBacΔF-HaF BV; 2, vHaBacΔF-SeF BV. (b) Western blot analysis of F proteins in infected cells. HzAm1 cells were infected with vHaBacΔF-HaF or vHaBacΔF-SeF at an m.o.i. of 10 TCID50 units per cell. SeUCR cells were infected with SeMNPV at the same m.o.i. as a control. At 72 h.p.i., cells were harvested for Western blot analyses. Lanes: M, protein molecular mass marker; 1, vHaBacΔF-HaF-infected HzAM1 cells hybridized with anti-HaF1; 2, vHaBacΔF-SeF-infected HzAM1 cells hybridized with anti-SeF1; 3, SeMNPV-infected SeUCR cells hybridized with anti-SeF1.

Western blot analysis was also used to detect the expression and cleavage of F proteins in infected cells (Fig. 5b). The results showed that, in vHaBacΔF-HaF-infected HzAM1 cells, a large amount of HaF was cleaved by furin, yielding a major F₁ subunit band of ∼59 kDa. There was still a minor band of uncleaved HaF₀ (80 kDa). In comparison, the cleavage of SeF in HzAM1 cells was not as complete, with only a small proportion of the SeF protein being processed into the F₁ subunit (∼60 kDa), leaving a large amount of uncleaved SeF₀ (80 kDa). This was quite different from the situation with SeMNPV in its host cells (SeUCR cells), where most of the SeF was cleaved (Fig. 5b). Interestingly, we observed that, although a large proportion of SeF remained uncleaved in HzAm1 cells, the F protein incorporated into the surface of pseudotyped vHaBacΔF-SeF BVs appeared to be cleaved (Fig. 5a).

Neutralization assays of pseudotyped viruses with specific and non-specific antisera
To test further the functionality of SeF relative to HaF, neutralization assays were carried out using monospecific polyclonal antibodies against SeF₁ and HaF₁ and using vHaBacΔF-HaF and vHaBacΔF-SeF (m.o.i. of 1 TCID₅₀ unit ml^–1) (Fig. 6). Anti-HaF₁ (Fig. 6a) and anti-SeF₁ (Fig. 6b) antisera specifically neutralized the infectivity of vHaBacΔF-HaF and vHaBacΔF-SeF, respectively, in a dose-dependent manner. The neutralizations were highly efficient, in that with 25 µl anti-HaF₁ antiserum or 1 µl anti-SeF₁ antiserum, more than 98 % infection by vHaBacΔF-HaF or more than 99 % by vHaBacΔF-SeF infection, respectively, was neutralized. The same amounts of pre-immune sera had no effect on virus infection compared with the virus-alone controls (data not shown). Anti-HaF₁ antiserum also showed some cross-neutralization of vHaBacΔF-SeF: 50 % inhibition of infection was achieved with 25 µl anti-HaF₁ antiserum (Fig. 6a). The results suggest that both SeF₁ and HaF₁ subunits contain neutralizing epitopes and that some of the epitopes may have similar structures.

(34K): Fig. 6. Neutralization assays of pseudotyped viruses. vHaBacΔF-HaF or vHaBacΔF-SeF was incubated with increasing amounts of anti-HaF1 (a) or anti-SeF1 (b) antibodies for 1 h at room temperature. The mixture of virus and antiserum was used to infect HzAM1 cells (m.o.i. of 1 TCID50 unit per cell). The infection rate of HaBacΔF-HaF (hatched bars) and HaBacΔF-SeF (open bars) were quantified at 60 h p.i by flow cytometry. Results are shown as means±SEM.

The F proteins of baculoviruses are structurally more diverse than the GP64 proteins (Pearson et al., 2000). For HaF and SeF, the amino acid identity is relatively low (∼33 %) and the question is whether this difference precludes the functionality of a heterologous F into group II NPVs. For this purpose, we successfully generated a group II F knockout bacmid (HaBacΔF; Fig. 1) and showed that F is essential for cell-to-cell transmission of infection. Our results also demonstrated that vHaBacΔF-SeF produced infectious BVs with kinetics similar to those of vHaBacΔF-HaF in HzAM1 cells, and that the fusion ability of these two pseudotyped viruses was similar. The successful substitution of HaF by SeF showed experimentally that F proteins are functional analogues in a group II NPV background. Despite low amino acid identity, the assembly of BVs was not compromised. This suggests a low degree of specificity of F in BV assembly in a group II NPV context. It has already been shown that F proteins of group II NPVs can pseudotype group I NPVs lacking GP64 (Lung et al., 2002). We have demonstrated here that type II F proteins can also pseudotype other F-null type II NPVs.

One-step growth curve analysis (Fig. 3) showed that the infectious BV production of vHaBacΔF-SeF was lower than that of vHaBacΔF-HaF, in which the F function was rescued. This observation might be explained by the incomplete cleavage of SeF in HzAM1 cells. The F protein is expressed as a precursor that undergoes cleavage by a pro-protein convertase (furin) of the host for activation (IJkel et al., 2000; Westenberg et al., 2002). In this respect, the processing of F proteins occurs in a fashion similar to the EFPs of Paramyxoviridae, Orthomyxoviridae, Togoviridae, Retroviridae and Herpesviridae (Lazarowitz et al., 1971; Meyer et al., 1990; Morse et al., 1992; Scheid & Choppin, 1977; Veronese et al., 1985). The furin enzyme is located in the trans-Golgi network and cleaves the EFPs in a virus-independent manner at the R-X-R/K-R motif (Hosaka et al., 1991; Vey et al., 1994). In vHaBacΔF-SeF-infected HzAM1 cells, furin cleavage of SeF was not as complete as cleavage of HaF in vHaBacΔF-HaF-infected cells (Fig. 5b). However, it was difficult to detect uncleaved SeF₀ present in the BVs of HaBacΔF-SeF (Fig. 5a).

It has been reported that in the human immunodeficiency virus (HIV) infection process, although uncleaved forms of gp160 (HIV-1) or gp140 (HIV-2) are extensively accessible at the cell surface, the precursor is not incorporated into virus particles following budding from the cell surface (Moulard et al., 1999). The high proportion of uncleaved gp160 was explained by a large amount of glycoprotein retained in the endoplasmic reticulum (ER), and only a minority of the gp160 molecules reach the trans-Golgi compartment for furin cleavage (Hallenberger et al., 1992). Therefore, the incomplete cleavage of SeF may be due to SeF being retained in the ER of HzAM1 cells with only relatively few molecules of SeF reaching the trans-Golgi to be cleaved. The incomplete cleavage of SeF may also be due to the fact that HzAM1 cells were used, which is not a normal host cell for SeMNPV. It is possible that the furin in HzAM1 cells recognizes the RNKR motif of HaF₁ more efficiently than the RSKR motif of SeF. As only cleaved SeF protein was incorporated into HaBacΔF-SeF BV particles, the lower titre of vHaBacΔF-SeF may be explained by a lower amount of correctly processed SeF proteins available on the cell membrane of HzAM1 cells for efficient viral budding. The different processing efficiency of the F proteins needs to be investigated further, and may reveal a process important for host-range determination. In this report, however, we did not investigate whether the lower infectious BV production of vHaBacΔF-SeF was due to lower BV production or to lower infectivity per BV. Further experiments such as real-time quantitative PCR will be carried out to answer this question.

Antisera against virus EFPs can neutralize infection. Therefore, we tested the neutralizing ability of polyclonal antibodies against HaF₁ and SeF₁. The F₁ subunit of F proteins includes the fusion peptide, the predicted heptad repeats, and the TM and CTD regions. The N-terminal peptide in F₁ and the heptad repeats have been proven to be functional as fusion and oligomerization regions, respectively (Long et al., 2006a; Westenberg et al., 2004). The antisera against SeF₁ and HaF₁ had specific neutralizing effects against the pseudotyped virus harbouring their respective F protein, suggesting there are neutralizing epitopes in F₁ subunits. A similar situation exists with paramyxoviruses and HIV-1, where several neutralizing epitopes in paramyxovirus F₁ domains and HIV-1 gp41 have been identified (Toyoda et al., 1988; Zolla-Pazner et al., 1999). The antiserum against the HaF₁ domain also had cross-neutralizing effects on vHaBacΔF-SeF, suggesting some conformational similarities between the two F proteins.

The results described in this paper suggest that the interactions of baculovirus F orthologues with other viral or host proteins are not species-specific, at least in the case of HaF and SeF. It remains to be seen whether this holds true for all F proteins or all group II viruses, in particular those that are more distantly related. It was suggested that the CTD of F, which was dispensable in pseudotyping AcMNPV, may be involved in a specific interaction with group II BV proteins (Long et al., 2006b). Our data suggest that these interacting proteins, if there are any, should be able to cross-interact with CTDs of different group II F proteins, at least in the case of SeMNPV and HearNPV. It has been shown that GP64 can be expressed on the surface of HearNPV BV and hence can promote the transduction of group II NPVs in mammalian cells (Liang et al., 2005). However, this insertion was in the presence of the homologous HearNPV F protein (Westenberg & Vlak, 2008). Recently, a report revealed that GP64 alone failed to pseudotype an F-null SeMNPV (Westenberg & Vlak, 2008). This can now also be tested experimentally through the availability of HaBacΔF (Fig. 1). Such investigations will help us to understand the similarity and differences between GP64 and F protein functioning.

The work was supported by NSFC grants (30470076, 30630002, 30670078), the 973 project (2003CB114202) and a PSA project from MOST and KNAW (2004CB720404). The authors would like to thank Dr Xiulian Sun for statistical analysis, Ms Yanfang Zhang for cell culture and Dr Basil M. Arif for scientific editing of the manuscript.