Baculovirus Host-Range

Reviews of literature on baculovirus host-specificity studies conducted up to the mid 1960s (57) and from the mid 1960s through the early 1980s (42) indicated that baculoviruses were only rarely transmitted across families and that the host-ranges of GVs were found to be narrower than NPVs. However, the earlier review cataloged reports of three successful transmissions across orders (57). It should be noted that the author of this review accepted the interpretations of host-specificity presented in the original papers. The subsequent application of molecular tools to characterize baculoviruses suggests that these early reports of cross-order infectivity are probably not valid.

The use of molecular identification methods for validating cross-infection is exemplified by an extensive study of Mamestra brassicae NPV (MbNPV) host-range (26). In this study DNA dot blots and restriction enzyme digests were used to verify infections. Although OBs were observed in Neodiprion sertifer (Hymenoptera: Symphyta) this apparent cross-order transmission was not confirmed as MbNPV by dot blot. No signs of infection were observed in other species in the orders Hymenoptera, Coleoptera, or Neuroptera that were tested. But MbNPV had a surprisingly broad hostrange in Lepidoptera particularly among Noctuidae although infection was restricted to three families. Moreover all infections outside of the Noctuidae required higher doses of virus to initiate infection.

AcMNPV and several closely related viruses infect the broadest range of insect species, although as observed for MbNPV (26), these species vary in their susceptibility to infection. In the Gr ner review Ac-MNPV was reported to infect 33 lepidopteran species in 10 families (42). A few other baculoviruses that are closely related to AcMNPV also exhibit relatively broad host-ranges. For example, Anagrapha falcifera NPV (AfNPV), later shown to be the previously described Rachiplusia ou MNPV (RoMNPV) (49), was found to infect 31 lepidopteran species in 10 families (56, 114). Several of these species had not been previously reported to be susceptible to AcMNPV, and one species, Helicoverpa zea, was significantly more susceptible to AfNPV than to AcMNPV (56, 114). In contrast, Bombyx mori NPV (BmNPV) another close relative of AcMNPV has a narrower host-range, reportedly infecting insects in seven lepidopteran families (42). Although comprehensive host-range studies validated by unequivocal virus identification are lacking for most baculoviruses, it appears that in general baculoviruses have relatively restricted host-ranges. Our current understanding is that baculovirus host-range is restricted to order and that within a single order, most are restricted to a single family and many to a single species or a few closely related species.

Virus host-range is determined by the ability of a virus to enter the cells of an organism, replicate its genome, assemble new virus particles, and release infectious progeny. Studies into these processes are facilitated by the availability of cultured cells that support the entire life cycle of the virus. For baculoviruses robust replication in cell culture has been achieved for NPVs infecting Lepidoptera (39, 54, 116, 121) and to a limited extent for GVs (40, 77, 98, 133). Although infection in cell culture can provide important insights into the molecular determinants of baculovirus host-range, it does not necessarily reflect a true organismal host-range. Virus host-range is less restricted in cultured cells and it is not uncommon for NPVs to replicate in cells derived from non-host species. Moreover, cell culture studies do not account for physical barriers to infection, immune defenses, or even limitations to infection due to tissue tropisms in an animal.

Understanding what determines the host-ranges of lepidopteran NPVs is further complicated by their bi-phasic lifecycles. During their lifecycle two morphological forms of virus are produced that have different functions and properties (8, 33, 34). Occlusion derived virus (ODV), embedded in polyhedral OBs, spreads the virus horizontally and is orally infectious for caterpillars. OBs are ingested by caterpillars and dissolve in the highly alkaline midgut releasing the enveloped virions. The virus envelope fuses with cellular membranes at the microvillar surfaces of midgut epithelial cells in a process that is poorly understood, but novel in that optimal fusion occurs at pH greater than eight (55). Following cell entry, nucleocapsids are transported to the nucleus where they uncoat and replicate. Budded virus (BV) is produced following the primary infection and spreads the virus to other cells within the infected caterpillar. Infection of tracheal cells facilitates BV movement across the basal lamina from the sites of primary infection in midgut epithelial cells to the haemocoel and to other tissues (30, 32). BV acquires its envelope from infected cell membranes and bears virally encoded fusion proteins, not found on the envelopes of ODV. BV enters cells by receptor mediated endocytosis (118, 119). The fusion proteins induce membrane fusion in the endosome at low pH, releasing the nucleocapsid (9, 58, 82, 101, 118, 119). BV are highly infectious for cultured cells and are used almost exclusively for host-range studies employing cell lines. ODV are much less infectious for cultured insect cells than BV (122), but are able to infect some cell lines (84).

The first critical step in a productive infection is entering a host cell. This involves virus binding to a cellular receptor. Therefore the presence or absence of an appropriate virus receptor is often a major determinant for virus host-range. In the case of NPVs little is known about the specific virus receptors. Only a few binding studies have been conducted on ODV. These studies suggest that ODV binds to specific receptors and that different viruses may have different receptors (46, 47, 55, 99). Specific binding of ODV to midgut receptors was first demonstrated for Lymantria dispar NPV (LdNPV) in binding studies using L. dispar brush border membrane vesicles (BBMV) (55). Subsequent to this study several NPV genes were identified that encode proteins required for oral infectivity. These per os infectivity factors (pifs) P74, pif1, pif2 and pif3 (72, 75, 99, 103) are all components of the ODV envelope and all but PIF3 are involved in receptor binding (46, 99, 109). Interestingly, oral infection with a pif-deletion mutant could not be rescued by co-infection with another virus bearing the gene, suggesting that all four proteins must be components of the same virus envelope to initiate infection (109). AcMNPV ODV from wt virus bound Heliothis virescens midgut epithelium more efficiently than ODV from p74 deletion viruses, but there were no differences in membrane fusion. The three-fold differences in binding efficiencies between the viruses could not account for the absence of oral infectivity by the P74 deletion virus, suggesting additional roles for P74 in initiating primary infection (46). AcMNPV is infectious for Spodoptera frugiperda, but only at high dosages most likely because of inefficient initiation of a primary infection (45). When the binding of AcMNPV and S. frugiperda NPV (SfMNPV) ODV to S. frugiperda midgut epithelial cells were compared, AcMNPV ODV bound at much lower levels that SfMNPV, but once bound there was little difference in membrane fusion between the viruses (47). AcMNPV did not compete with SfMNPV for binding in co-inoculation studies with labeled virus suggesting the two viruses bind to different receptors. These studies implicate specific midgut receptors for oral infection by NPV, but it is still unknown if they have a role in determining baculovirus host-range. Indeed, the p74 deletion virus study suggests that events subsequent to binding may be more important for initiating a productive infection (46). One report suggested that P74 might be a host-range determinant because replacement of the AcMNPV p74 gene with the Spodoptera litura multicapsid nucleopolyhedrovirus (SpltMNPV) p74 gene abolished AcMNPV oral infectivity for Argyrogramma agnata larvae (134). However a role for P74 in host-specificity has not been independently confirmed and there may be other explanations for these results, for example, it is possible that SpltMNPV P74 is incompatible with the other AcMNPV pifs (109).To date there have been no comparative studies of ODV binding of an NPV to midgut epithelial cells from insect species with different levels of resistance to oral infection. Identification of pif genes involved in virus attachment should facilitate the identification of specific receptors for ODV to help resolve these questions.

Although a specific receptor has yet to be identified, AcMNPV BV can enter diverse cells, including mammalian cells, and express a reporter gene (11, 23, 93, 94). These studies indicate that the receptor for BV is highly conserved and the block to baculovirus infection in cultured cells, at least for AcMNPV, occurs subsequent to entry and transport to the nucleus. While these observations indicate that BV receptor binding has little influence on baculovirus specificity at the cellular level, there is a complication. Two types of fusion proteins are associated with the BVs, GP64 and F-protein. AcMNPV BV bears GP64 (117, 120). F-protein is the ancestral fusion protein, whereas GP64 was acquired more recently and is related to the fusion proteins of Thogotovirus, an orthomyxovirus (58, 96, 101, 102). The GP64 type of fusion protein is used by a subset of baculoviruses, termed group Ⅰ NPVs (52, 137), which are phylogenetically related to AcMNPV. The others, the group Ⅱ NPVs, bear F-protein on the BV envelope. Both GP64 and F-proteins promote membrane fusion in a similar manner and replacing AcMNPV gp64 with F-proteins from LdMNPV, S. exigua NPV (SeNPV), or Heliothis armigera NPV (HearNPV) resulted in pseudotyped AcMNPV BV that could initiate membrane fusion in Sf9 cells (78, 82). Competitive binding studies between group Ⅰ and group Ⅱ NPVs demonstrated that they bind to different receptors (130, 131). However, the ability of AcMNPV pseudotyped with different F-proteins to fuse with cells that are normal hosts for AcMNPV but not hosts for LdMNPV or HearNPV, the sources of the F-protein genes, suggests that specific receptor binding does not play a key role in restricting the host-ranges of type Ⅱ viruses.

While it is tempting to rule out viral entry as a factor affecting baculovirus host-range, the ability of BV from few baculoviruses, other than AcMNPV, to bind to and enter diverse cell lines have been investigated using sensitive techniques such as reporter genes. One such study implicates GP64 as a host-range factor for BmNPV (71). As noted above, AcMNPV has a broader host-range than BmNPV, yet they are closely related group Ⅰ NPVs, with common genes that share over 90% nucleotide and amino acid identity (35). The difference in host-range extends to cultured cells, making AcMNPV, BmNPV, and the cells that support their replication ideal models for investigating the molecular basis of baculovirus host-range. This study investigated the ability of BmNPV to infect seven non-permissive or semi-permissive insect cell lines, using a virus expressing a fluorescent reporter gene driven by the ie-1 promoter (71). Fluorescence was observed in only a small percentage of infected Sf9 and BTI-Tn 5B1-4 (TnHi5] cells, but the percentage of fluorescing cells increased and detectable viral progeny were produced when viral DNA was transfected into cells. Localization of the virus via immuno-cytochemistry showed that the virus trafficked to the region surrounding the nucleus, but failed to enter it in the majority of infected cells. When the BmNPV GP64 was replaced with AcMNPV GP64, the pseudotyped BmNPV could then enter the nucleus and complete its lifecycle in TnHi5 cells, but not in Sf9 cells. This suggested that the BmNPV GP64 could not efficiently mediate fusion in the endosome to release the nucleocapsids, raising the possibility that fusion proteins may indeed have a role in host-range, albeit subsequent to cellular entry. Previous studies that examined BmNPV infection in Sf9, Sf21, and Tn368 cells found either limited infection, or minimal DNA replication and an extremely low level of progeny virus produced, consistent with productive infection in only a small subset of infected cells (86), as was observed in this study. These results are also consistent with the low levels of virus gene transcription observed in Sf9 cells infected with BmNPV or transfected with BmNPV DNA (65). A puzzling aspect about this study arises when comparing it with several earlier studies on BmNPV and AcMNPV host-range. Specifically, it is hard to reconcile the inability of BmNPV to enter the nuclei of Sf9 with studies in which recombination between AcMNPV and BmNPV occurred in Sf21 cells since it would require both viral genomes in the nucleus (73). See discussion of p143, below. One possibility is that Sf9 and Sf21 differ in this respect and in these co-infection studies AcMNPV provided a “helper function”, which allowed BmNPV to enter Sf9 cells. This possibility is consistent with the study of Rahman and Gopinathan in which they observed no BmNPV replication in BmNPV-infected Sf9 cells unless they were co-infected with AcMNPV (105). The Katou et al. study implicates AcMNPV gp64 as providing the putative “helper function” (71). AcMNPV and BmNPV gp64 proteins are highly conserved, sharing 95% amino acid identity. However, they reported more efficient membrane protein fusion by AcMNPV gp64 than BmNPV gp64 in syncia forming assays. But it is unclear how gp64 could facilitate nuclear entry in trans, in the co-infection studies (105). This study also illustrates the diversity amongst the numerous cell lines support AcMNPV replication. When the block to nuclear entry was relieved by pseudotyping BmNPV with AcMNPV gp64, BmNPV could replicate efficiently in TnHi5 but not Sf9 cells, indicating another block to infection. The additional block for infection in Sf9 cells may be related to trafficking and function of p143, p35, or both in these cells (5). See below.

With the exception of the study reported above, most available evidence indicates that the restriction to baculovirus replication in cultured insect cells occurs subsequent to virus entry. However, infection is blocked at different steps in the replication cycle depending on the specific combination of virus and cell line (44, 71, 86, 88, 94, 123). These steps, that include transport to the site of replication, viral gene expression, and generation of viral progeny, require complex interactions between the virus and its host cell. Hundreds of cell lines have been established from lepidopteran cells, (37, 38, 83) and have the potential to provide a wealth of knowledge regarding the molecular bases of baculovirus host-range. However this resource is largely untapped. As described in the following paragraphs, studying the interactions of baculoviruses in a few of these cell lines has identified six genes that affect host-range. Several of these were confirmed to alter baculovirus host-range in insects.

Studies of NPV replication in cultured cells identified p35, the viral DNA helicase p143, lef7, ie-2, hcf-1, and hrf-1 as virus genes that affect baculovirus host-range or tissue tropism. The majority of these were identified in AcMNPV. Among these, p143 is a core gene found in all baculoviruses that is required for viral DNA replication. The rest are auxiliary genes found in only a few baculoviruses or groups of baculoviruses. All but hrf-1 encode AcMNPV late expression factors (lefs), proteins required for late gene expression with roles in late gene transcription, DNA replication, or both (80, 107, 113). In addition to changing host-range in cultured cells, changes in the AcMNPV p143 coding sequence or insertion of the hrf-1 gene into the AcMNPV genome expanded AcMNPV host-range in insects. In contrast, deleting p35, lef7, or hcf-1 reduced the host-range of AcMNPV in cultured cells. Deleting p35 also reduced AcMNPV host-range in insects. These genes appear have very different functions, but with the exception of p35, our understanding of how the products of these genes affect host-range is still limited. The effects of these genes on baculovirus host-range were reviewed previously (90, 110), therefore we will briefly recap what was previously reported and present the results of subsequent research. In addition, we introduce another example of a baculovirus cell culture system currently under investigation that may reveal additional baculovirus host-range factor genes. These co-infection studies of Thysanoplusia orichalcea M NPV (ThorMNPV) and AcMNPV suggest that AcMNPV encodes a factor, or alternatively induces expression of a host factor, that expands the host-range of ThorMNPV.

The viral DNA helicase gene, p143, was the first baculovirus gene to be associated with virus host-range. The involvement of p143 in host-range was revealed when BmNPV and AcMNPV were used to co-infect TN368 cells and Sf21 cells, which support replication of AcMNPV but not BmNPV. Recombinant viruses arose that were also infectious for BmN cells, a cell line that does not normally support productive AcMNPV replication (73). These viruses had broader host-range in cultured cells, replicating in BmN as well as in Tn368 and Sf21 cells. They were also infectious for B. mori larvae when BV was injected and when larvae were orally infected with OBs (73). Repeated backcrossing of recombinants, by co-infecting with AcMNPV, resulted in a virus nearly identical to AcMNPV, called eh2-AcNPV that maintained the expanded host-range phenotype for BmN cells. These viruses had a small insert of BmNPV genomic DNA that mapped to a 600 bp region of p143 (85). A role for p143 in AcMNPV host-range expansion in cultured cells and larvae was confirmed independently with different B. mori cell lines and a different isolate of AcMNPV (2, 24). In these studies, recombinant viruses were isolated following co-transfection of Sf9 cells with AcMNPV DNA and BmNPV restriction fragments. These studies further refined the requisite changes in AcMNPV p143 amino acid sequence, to that of BmNPV p143 sequence, to three residues, Leu556, Asn564 and Leu577 for infection of cultured B. mori cells (24). The latter two amino acid changes (Asn564 and Leu577] were sufficient for AcMNPV to infect and kill B. mori larvae when BV was injected (2). These changes in the amino acid sequence of AcMNPV p143 were thought to reduce cytotoxicity caused by AcMNPV replication in the B. mori cells (68). When BmN cells are infected with AcMNPV, or co-infected with BmNPV and AcMNPV, a cytopathic effect is observed in which translation of both host and virus genes are attenuated. This effect was not observed when the cells were co-infected with BmNPV and eh2-AcNPV. It should be noted that not all isolates of AcMNPV induce the same degree of cytopathic effect in BmN cells. For example the AcMNPV L1 isolate did not inhibit BmNPV replication in co-infection experiments (68). It is possible that the observed cytotoxicity is the result of aberrant DNA replication or other activities of p143, as p143 is expressed at higher levels in AcMNPV-infected BmN cells than in Sf9 cells (65). Although the changes in AcMNPV p143 sequence allow AcMNPV to replicate in B. mori cells and larvae as well as in Sf9 cells when infected at high multiplicity of infection (moi), the altered p143 may not be fully functional in Sf9 cells since eh2-AcMNPV infection aborts at an early stage in the majority of cells when infected at a low moi (69).

BmNPV does not replicate in Sf9 or Sf21 cells, but it does not cause the type of cytotoxic effect that is observed when AcMNPV infects BmN cells (73, 86, 105). It is likely that the lack of cytotoxicity is because of low virus activity. BmNPV genes are expressed very poorly in Sf9 and Sf21 cells and there is little evidence of DNA replication (65, 73, 86, 105). These results are consistent with the findings of Katou et al. that indicated that transport of BmNPV to the nucleus was blocked in most but not all Sf9 cells. Yet, BmNPV failed to replicate efficiently in Sf9 cells even when the problem of transport was solved (71). Given the ability of a few amino acid changes in AcMNPV p143 to expand its host-range to B. mori larvae and cell lines, might problems with BmNPV p143 function be responsible for its inability to replicate in the Sf9 cells? A study that examined the ability of BmNPV lefs to functionally substitute for AcMNPV lefs in transient assays suggests this is the case (5). Nineteen AcMNPV lef genes were identified using a transient assay in which cloned fragments of AcMNPV genomic DNA were assayed for their ability to activate expression of a reporter gene driven by a late AcMNPV promoter (100, 113). In this study BmNPV p143 was one of only three BmNPV lefs, out of the 19 lefs, that could not functionally substitute for its AcMNPV counterpart in Sf21 cells in this assay (5). The others ie-2 and p35 have also been implicated in baculovirus host-range (see below). They found that the BmNPV p143 protein was not transported to the nucleus. Because lef-3 has been shown to interact with p143 and facilitate p143 nuclear transport (135), plasmids expressing Bm p143 and Bm lef3 were substituted together. Although Bm lef3 alone successfully replaced Ac lef3 in the transient expression assay, co-transfecting Bm lef3 and Bm p143 abolished reporter gene expression and Bm p143 was not transported to the nucleus. Mutating the AcMNPV p143 sequence to the BmNPV p143 sequence from Ser to Asn at residue 564 or Phe to Leu at 577 (2, 24), individually, restored late gene expression when the altered AcMNPV p143 was expressed with either AcMNPV or BmNPV lef3 (5). However, AcMNPV p143 with both mutations had little activity when expressed with either AcMNPV or BmNPV lef3, and less with BmNPN lef3 than AcMNPV lef3. AcMNPV p143 with the double mutation physically associated with both BmNPV and AcMNPV lef3 and localized to the nucleus. Bm p143 with reciprocal mutations, to the AcMNPV sequence, e.g., Asn to Ser at residue 564 or Leu to Phe at 577, or at both residues, were unable to substitute for AcMNPV p143, with either BmNPV or AcMNPV lef3, to activate reporter gene expression. Moreover, the modified BmNPV p143 did not physically interact with either lef3 and did not localized to the nucleus. No reporter gene activity was observed when Sf21 cells were transfected with all BmNPV lefs, and AcMNPV p35. Low level reporter gene expression was observed when BmNPV p143 was replaced with AcMNPV p143 in these assays. These observations support the idea that the inability of BmNPV p143 to properly function in Sf21 cells contributes to, but is probably not the sole reason, for the inability of BmNPV to replicate in Sf21 cells.

AcMNPV p35 functions as a host-range factor because deleting the p35 gene reduces the ability of AcMNPV to infect Sf21 cells (20, 53) and S. frugi-perda larvae (21, 22). For larval infections the route of infection was important. S. frugiperda larvae were 1000 fold less susceptible to △p35AcMNPV ODV delivered orally via OBs than to injected BV (22). This may be due to elimination of infected midgut cells during primary infection before the virus can spread systemically (124, 126, 128, 138). The ability of △p35AcMNPV to infect Sf21 cells was restored when it carried a different apoptotic suppressor gene, inhibitor of apoptosis (iap) from the Cydia pomonella GV (25), suggesting that the role of p35 in AcMNPV host-range is preventing virus-induced apoptosis in S. frugiperda cells and larvae. When tissues of Ac-MNPV-infected S. frugiperda larvae were analyzed for apoptosis by TUNEL assay, those injected with △p35AcMNPV showed higher numbers of TUNEL positive cells relative to a revertant virus, further supporting this idea (18). Similar studies have not been done using an oral route of infection. Few other baculoviruses carry p35 or the related apoptotic suppressor p49 (27, 139), but most carry one or more copies of iap. Among the 48 baculoviruses, of all genera, sequenced to date seven carry p35 and four carry p49. Possible roles for the various apoptotic suppressors in virus host-range have not been extensively studied. P35 may be one of the factors that contribute towards the narrow host-range of BmNPV relative to AcMNPV as BmNPV p35 does not appear to be functional in Sf21 cells (5), was not essential for BmNPV replication in BmN cells or larvae (36, 70), and is less effective than AcMNPV p35 in blocking apoptosis in mammalian cells (92). The roles of p35 and two iaps carried by AcMNPV were tested in six cell lines, Sf21, Tn368 and four additional cell lines. Deleting both iaps had no effect on AcMNPV infection in any cell line. Whereas deleting p35 resulted in apoptosis in AcMNPV-infected Ld652Y and Sl2 cells, two cell lines that do not support AcMNPV replication, and had no effect in M. brassicae or Panolis flammea cells (41). Beyond these reports, to our knowledge, roles for the apoptotic suppressors of other baculoviruses in determining their host-ranges have not been investigated.

Lef7 genes have been found only in baculoviruses and were identified in ten out of the 44 fully sequenced baculoviruses, all of them type Ⅰ NPVs. Lef7 was identified in transient assay screens for AcMNPV genes required for enhancing late gene expression in Sf21 cells (95). However when the lefs required for late gene expression in Sf21 cells were assayed in Tn368 cells lef7 was non-essential, as were p35, and ie-2 (79). Further support for the role of lef7 as a host-range factor was provided when deletion of lef7 was identified as the defect in two AcMNPV mutants that were defective for AcMNPV replication in Sf21 cells but not Tn368 cells (14). Sf21 cells infected with AcMNPV lacking lef7 have dramatically reduced titers, produce few if any OBs, and DNA replication is barely detectable (14). The effects of lef7 deletion on AcMNPV infectivity in S. frugiperda and T. ni larvae have not been investigated. Lef7 is also non-essential for BmNPV replication in BmN cells (36). In BmN cells infected with lef7 mutated BmNPV, viral DNA replication was reduced and late transcription was enhanced relative to wt BmNPV, but there was no noticeable effect on BV production. Moreover, deleting lef7 did not impair BmNPV infectivity for B. mori larvae when BV was injected (36). Together these studies suggest that lef7 functions as a cell or tissue specific factor required either directly or indirectly for viral DNA replication.

Hcf-1 is a gene from AcMNPV that is necessary for it to infect Trichoplusia ni cell lines (79, 81). When Tn368 cells were transfected with the AcMNPV lef library in transient expression assays as described above, reporter gene expression was weak. This suggested that additional lefs were required for AcMNPV late gene expression in Tn368 cells. Further screening revealed a single additional gene required for gene expression in Tn368 cells. This gene corresponded to AcMNPV ORF70 and was renamed host cell factor 1, hcf-1 (79). Hcf-1 has no homologues outside of baculovirus, and a paucity of motifs that might suggest a function. The predicted protein is rich in cysteine residues and has a putative C4C4 RING finger motif (1, 48). RING motifs coordinate zinc ions and are associated with various cellular functions including, protein-protein interaction and ubiquitin ligase activity (10, 67, 108). Interestingly, hcf-1 is one of only three lef genes found in the AcMNPV genome, but missing from the BmNPV genome (35). In fact hcf-1 has been identified in only 4 of the 44 completely sequenced baculovirus genomes, AcMNPV (L22858], RoMNPV (AY145471], Plutella xylostella NPV (PlxyNPV) (DQ457003], and Clanis bilineata NPV (ClbiNPV) (DQ504428]. PlxyNPV and RoMNPV are type Ⅰ NPVs, closely related to AcMNPV, and hcf-1 shares 99 and 84% amino acid identity and 99 and 91% similarity with AcMNPV hcf-1, respectively. In contrast, ClbiNPV is a type Ⅱ NPV and shares only 24% identity, and 43% similarity with AcMNPV hcf-1.

AcMNPV lacking hcf-1 replicates in Sf21 cells but its replication is impaired in T. ni cell lines (81). However, the effects are much more severe in Tn368 cells than in TnHi5 cells. In TnHi5 cells production of OBs were delayed by approximately 24 h and although OBs were observed in most cells by 72 h pi, they differed morphologically from those in wt AcMNPV-infected cells. In contrast, no OBs were observed in Tn368 cells infected with hcf-1 deletion virus. A slight increase in BV titer was observed over time in Tn368 cells infected by an hcf-1 deletion mutant, but translation of virus and host genes arrested by 18 h pi. Moreover viral DNA replication was barely detectable and the viral capsid protein could not be detected on Western blots. Deleting hcf-1 had no effect on AcMNPV oral infectivity for T. ni larvae, although it did reduce its virulence. However when hcf-1 mutant virus BV was injected there was a significant reduction in infectivity for both S. frugiperda and T. ni larvae, as well as reduced virulence in T. ni, relative to wt AcMNPV. Thus hcf-1 functions more like a virulence factor than a host-range determinant for AcMNPV in T. ni larvae, and appears to have an important role in cell or tissue tropism within the insect. Whether it might have a more definitive effect on larval host-range in another host species has not been determined. It would be interesting to determine if the addition of hcf-1 to the BmNPV genome would affect its ability to infect T. ni cells. Although the specific role for hcf-1 in infection is still unclear, its activity appears to require self-association and nuclear localization (51). Hcf-1 nuclear staining is diffuse in transfected cells but shows distinct punctate foci in infected cells, suggesting that it may localize to promyelocytic leukemia (PML) nuclear bodies (132). This suggests it could play a role in viral DNA replication or possibly immune response (4, 31). AcMNPV ie-2 and PE38, immediate early transcription factors also localize to these nuclear regions in transfected Sf21 cells (74), and associate with PML in mammalian cells (97). Like hcf-1, ie-2 and PE38 have RING fingers but they are not the C4C4 type found in hcf-1. Mutational analysis of hcf-1 function in AcMNPV infection implicates the RING-finger domain in its ability to support virus replication and to self-associate (132).

AcMNPV ie-2, also referred to as ie-N, is a transcription factor that stimulates transcription by the essential early transcription factor ie-1 (12). It was identified as an AcMNPV lef gene in Sf21 cells in transient expression assays where it stimulates late gene expression and plays a role in viral DNA replication (80, 100). In contrast to Sf21 cells, ie-2 was not necessary for late gene expression in Tn368 or Hi5 cells in transient expression assays, suggesting it was a host-range gene (79). The ie-2 gene is found in approximately one quarter of sequenced baculovirus genomes, all type Ⅰ NPVs. It has no known homologues outside of the baculoviruses and is characterized by the presence of a RING finger. The RING finger confers the ability to arrest the cell cycle in Sf21 and Tn368 cells, but is not involved in its ability to stimulate transcription (104). Mutant Ac-MNPV in which ie-2 is deleted or mutated in domains involved in either transcriptional activation or cell cycle arrest were viable in Sf21, Tn368, and TnHi5 cells. All of these mutants exhibited some level of deficiency in DNA replication, transcription, or production of BV and OBs in Sf21, Tn368, and TnHi5 cells, although the deficiencies were more severe in Sf21 cells than either T. ni cell line. All of these ie-2 mutant AcMNPV viruses exhibited reduced oral infectivity against S. frugiperda and T. ni larvae, but no differences in infectivity, relative to wt AcMNPV, when BV was injected. The reduction in oral infectivity could be attributed to a reduction in the number of ODV imbedded in OBs produced by mutant virus infections, rather than to functional deficiencies of ie-2 protein. BmNPV ie-2 mutants were also viable and produced normal levels of BV, however DNA replication was delayed and reduced relative to wt virus (36). These studies suggest that ie-2 is somewhat more important for AcMNPV replication in Sf21 cells than in T. ni cell lines. These cell line specific effects suggest that it may be involved in cell or tissue tropism, but it does not appear to play a significant role in baculovirus host-range. Based on the BmNPV studies, the defect in infections by ie-2 deletion mutants is reduced viral DNA replication. How ie-2 influences the number of ODV embedded in polyhedra is unknown, but it could simply be due to a reduction in the number of enveloped virions present in the nucleus during the occlusion process.

A baculovirus gene, hrf-1, was identified in a genetic screen for baculovirus host-range factors from L. dispar M NPV (LdMNPV) (112). This screen was based on the observation that AcMNPV progeny are not produced in Ld652Y, but the block to AcMNPV replication could be relieved by super-infecting AcMNPV-infected Ld652Y cells with LdMNPV (87, 88). The hrf-1 gene encodes a 26 kDa protein with no known motifs to suggest its function (112). To date, only one hrf-1 homolog has been identified and it is found in another baculovirus, Orgyia pseudotsugata M NPV (OpMNPV), which also replicates in Ld652Y cells. OpMNPV encodes a carboxy-terminal-truncated version of the LdMNPV hrf-1 with low sequence identity. It is 78 amino acids long compared to the 218 amino acids comprising LdMNPV hrf-1 (59, 112). AcMNPV infection of Ld652Y cells induces a cytopathic effect and all temporal classes of virus genes are transcribed, viral DNA replication is observed, but translation of both virus and host proteins arrests by approximately 12 h post infection (44, 88, 94). Ac-MNPV recombinant viruses bearing hrf-1 productively infect Ld652Y cells and translation arrest is rescued (28, 112). Moreover AcMNPV bearing hrf-1 was orally infectious L. dispar larvae, which are not susceptible to AcMNPV infection (13). In contrast, the OpMNPV hrf-1 homolog did not rescue translation arrest or support AcMNPV replication in Ld652Y cells. Ac-MNPV bearing a chimeric hrf-1 comprised of the OpMNPV hrf-1 fused to the carboxy-terminal coding region of the LdMNPV hrf-1 also failed to replicate in Ld652Y cells or rescue translation arrest (59). Because the OpMNPV hrf-1 is not functional in the context of AcMNPV-infected Ld652Y cells it is unclear if it is a functional homolog of LdMNPV hrf-1. However, it could play a role in the ability of OpMNPV to replicate in Ld652Y cells.

Neither the cause of translation arrest in Ld652Y cells, nor the mechanisms by which hrf-1 rescues translation arrest and expands the host-range of AcMNPV are understood. Hrf-1 does not function as an apoptotic suppressor. AcMNPV p35 is expressed and functions to prevent apoptosis in AcMNPV-infected Ld652Y cells (29). Hrf-1 could not prevent apoptosis in the absence of p35, although it did prevent global translation arrest in apoptotic cells prior to their final destruction (29). There does, however, appear to be a connection between p35 activity and translation arrest. In recombinant AcMNPV-infected Ld652Y cells, the presence of p35 in the absence of hrf-1 correlated with global translation arrest, suggesting that the activity of p35 might trigger translation arrest. In cells treated with the DNA synthesis inhibitor aphidicolin, translation is attenuated rather than fully arrested, but strikingly, is only observed in Ld652Y cells infected with AcMNPV expressing p35 but not hrf-1 (29, 111). Infection with a AcMNPV p143 mutant, a temperature sensitive replication mutant, had a similar effect as aphidicolin on translation arrest in wt AcMNPV-infected Ld652Y cells (111). These two observations suggest that early activity of p35 plays some role in initiating translation arrest, but an event associated with DNA replication or other event in the late stage of the virus replication is what drives global translation arrest. Apoptosis in Sf21 cells infected with p35 deletion mutants exhibits an identical pattern, early initiation but a dramatic increase at the onset of DNA replication in the late stage of infection (76).

The anti-apoptotic function of p35 clearly plays a role in inducing translation arrest either directly or indirectly (29, 111). Infecting Ld652Y cells with AcMNPV expressing p35 with functional mutations that do not block apoptosis (6), does not induce global translation arrest (6, 111). Moreover, AcMNPV expressing other baculovirus apoptotic suppressors, OpMNPV iap3, CpGV iap3, or SlNPV p49, in lieu of p35 induce global translation arrest in a manner identical to p35. Surprisingly, treating cells with peptide caspase inhibitors to block apoptosis in Ld652Y cells infected with the AcMNPV p35 deletion mutant did not result in translation arrest. The inability of peptide inhibitors of apoptosis to induce translation arrest in AcMNPV-infected Ld652Y cells, while effectively inhibiting apoptosis, suggests that baculovirus apoptotic suppressors have additional functions beyond simply preventing apoptosis. The nature of this function is puzzling since two very different types of apoptotic suppressor proteins, P35-like and IAP, that prevent apoptosis by different mechanisms (19), have the same effect on translation in the AcMNPV-infected Ld652Y cells. One possibility is that the viral proteins not only prevent apoptosis, but stimulate viral DNA replication, late gene transcription, or other activity late in the virus replication cycle, in a way that the caspase inhibitors cannot fully mimic. These late events in the virus replication cycle might then in turn drive or enhance the translation arrest mechanism. Given the differences amongst the different types of baculovirus apoptotic suppressors, it is unclear how each might stimulate late events in the AcMNPV replication cycle.

Mutagenesis of hrf-1 provided few clues to its function (59). Truncating the protein at either the amino-or the carboxyl-termini of the protein and two-amino acid insertional mutations throughout the length of the protein abolished its ability to rescue translation arrest. As noted above chimeric hrf-1 in which the amino terminus was replaced with hrf-1 from OpMNPV did not prevent translation arrest, suggesting the amino terminus of hrf-1 is important for its function. The mutant proteins also appeared to be unstable in mutant AcMNPV-infected Ld652Y cells, but not in Sf21 cells. Although lacking any characteristic functional motifs, hrf-1 possesses one highly acidic domain. Some of the mutations in this region reduced or abolished translation arrest whereas others had little effect. These results suggested the structure of hrf-1 was important for its function and stability and that the acidic region might play a functional role in rescuing translation arrest. Reasons for the differences in expression and/or stability of hrf-1 mutant proteins in Ld652Y relative to Sf21 cells are unknown.

Regardless of how hrf-1 functions it appears to be a critical virus factor for infecting Ld652Y cells. Although Ld652Y cells are permissive for LdMNPV and Op-MNPV replication, they do not support the replication of a number of other baculoviruses. In screens of Ld652Y cells with seven different baculoviruses, including AcMNPV, apoptosis was induced to some degree by Hyphantia cunea NPV (HycuNPV), Spodop-tera exigua NPV (SeMNPV), Spodoptera littoralis NPV (SpltNPV), BmNPV, and OpMNPV (63). Apoptosis was most pronounced in HycuNPV-infected cells and the least in the OpMNPV-infected cells. In a subsequent study treating HycuNPV-infected Ld652Y cells with peptide apoptotic suppressors prevented apoptosis, but progeny virus was not produced (62). As in AcMNPV-infected Ld652Y cells, translation arrested in HycuNPV-infected Ld652Y cells by 12 h pi. However, when Ld652Y cells were infected with a recombinant HycuNPV bearing hrf-1 progeny BV were produced and translation arrest was prevented. Even with the addition of peptide caspase inhibitors no OBs were observed and no polyhedrin protein was produced in Ld652Y cells infected with HycuNPV bearing hrf-1. No virus progeny were produced in BmNPV-, SeMNPV-, or SpltNPV-infected Ld652Y cells. Expressing hrf-1 from a transfected plasmid promoted productive BmNPV and SeMNPV, but not SpltNPV, infections. In contrast to HycuNPV-infected cells, both BV and OBs were produced. This study suggested that translation arrest could be a cellular mechanism for limiting baculovirus infection by Ld652Y cells. In the case of HycuNPV there is an additional block to a very late events in infection preventing OB production, but the nature of that block is currently unknown. Surprisingly, when hrf-1 was incorporated into BmNPV genome, it was less effective in promoting BmNPV replication in Ld652Y cells than hrf-1 supplied in trans (64). The reason for this difference is unclear, but could possibly be due to lower levels of hrf-1 expression in the context of virus infection, because pre-transfecting hrf-1-expressing plasmids increased the expression of late virus proteins.

Thysanoplusia orichalcea M NPV (ThorMNPV) is a recently discovered NPV that exhibits a relatively narrow host-range in insect cell lines and larvae (17, 123). Phylogenetic analysis using polh, lef-8 and lef-9 genes predicted that ThorMNPV is clustered with AcMNPV in the group Ⅰ NPV (17). ThorMNPV replicates in TnHi5 cells very well and high BV titers can be obtained (17, 123). Using a recombinant Thor-MNPV containing a GFP-reporter controlled by the AcMNPV polyhedrin promoter, vThGFP, eight of 28 cell lines tested were permissive at different degrees for vThGFP as determined by green fluorescence (123). In contrast to TnHi5 cells in which over 95% of cells fluoresced and produced OBs by 3 days p.i., only a few Sf21 cells ( < 1%) were infected by 3 days p.i.

Co-infection studies of vThGFP and vAcRed, a recombinant AcMNPV expressing the red fluorescent protein (RFP) gene from the polyhedrin promoter, in Sf21 cells demonstrated that AcMNPV could promote ThorMNPV replication in the Sf21 cells. In the co-infected cells, 65% of the cells showed only green fluorescence indicating ThGFP replication, 33% of the cells showed red fluorescence indicating vAcRed replication, and about 2% of cells showed yellow when GFP and RFP channel images were merged, indicating replication of both vThGFP and vAcRed in the same cells. OBs were formed in majority of cells expressing GFP by vThGFP. It was surprising that so few cells expressed both GFP and RFP, indicating that replication of one virus inhibited replication of the other virus (16, 123).

There are at least three possible explanations for the enhanced infectivity of vThGFP by vAcRed in Sf21 cells: recombination between the viruses, an AcMNPV gene product acting in trans, or secretion of a cellular (or virus) factor from the AcMNPV-infected cells. Enhanced infectivity of vThGFP does not appear to be the result of recombination with vAcRed, since no recombinant virus could be isolated from co-infected Sf21cells (16, 123). The possibility that AcRed provided trans-activation factors that enhanced vThGFP infection in Sf21 also appears unlikely because co-transfecting Sf21 cells with an AcMNPV cosmid library failed to identify any gene(s) that could enhance vThGFP infection of Sf21 cells. However, centrifuged media from AcMNPV-infected Sf21 cells enhanced vThGFP infection of Sf21 cells 15-fold (123), supporting the third hypothesis. At the present time the secreted factor from AcMNPV-infected Sf21 cells has not been identified.

At the organismal level, ThorMNPV is highly virulent to three species, T. orichalcea, T. ni, and Pseudoplusia includens. Other insects such as S. frugi-perda, S. exigua, and H. zea are resistant to Thor-MNPV (17). When the larvae of resistant species were fed vThGFP OBs, no mortality was observed and all larvae developed into pupa and then adults. However, hemocytes, trachea and Malpighian tubules showed GFP fluorescence demonstrating secondary spread of infection from the midgut, although no OBs were observed in any of these tissues. GFP expression was not detected in fat bodies from any of the three resistant insects and may be the reason oral infection with vThGFP OBs could not kill these hosts because fat body is the major tissue of NPV replication in native hosts. Levels of infection in the other tissues where fluorescence was observed, hemocytes, tracheal and Malpighian tubules, may not be sufficient to kill the insect or impede development (123). Intrahemocoelic injection of vThGFP BVs to these nonsusceptible insects (S. frugiperda, S. exigua and H. zea) produced high mortality, which suggested that although BV was produced in the midgut in orally infected insects and was able to spread to a limited number of tissues, the levels of BV were not sufficient to initiate a robust secondary infection (123).

All of the host-range genes identified, to date, were isolated functionally using virus infected insect cells in culture. Each relied on the availability of systems of baculoviruses and cell lines with suitable properties for identifying a host-range determinant, such as the closely related AcMNPV and BmNPV with distinctly different host-ranges in cultured cells. With advances in rapid DNA sequencing technology coupled with decreasing costs, genomic analysis promises to be an important and powerful tool for increased understanding of how baculovirus host-range is determined and for identifying the genes responsible (52, 115). If the currently known host-range genes are reasonable examples, additional baculovirus host-range genes are likely to be found amongst the unique or less widely distributed auxiliary baculovirus genes.

Understanding the role of insect immunity in restricting baculovirus host-range at the organismal level is among the most intriguing questions still remaining. Research to date suggests that the relative ability of the host insect to resist or eliminate virus infection plays an important role in determining baculovirus host-range. In vivo studies using marked viruses and immune suppressors demonstrated that highly resistant species could be infected if their immune systems were compromised (125, 127). Genetic studies with susceptible and resistant B. mori strains reveal a dominant gene or set of linked genes to be responsible for B. mori resistance to AcMNPV infection, but are not yet identified or cloned (43). In this study, AcMNPV BV was injected into the larvae, so it is also not clear whether or not the gene(s) affect B. mori resistance to AcMNPV oral infection. It is also not known if B. mori resistance to AcMNPV is part of an immune response. As discussed above, cellular responses such as apoptosis and translation arrest can prevent or reduce productive baculovirus infection in a cell-line-specific manner. But how the cell senses that it is infected and signals for a response is unknown. Both responses also appear to be important at the organismal level. Apoptosis is considered an immune response against virus infection and at least for AcMNPV, the ability to prevent apoptosis in S. frugiperda due to the activity of p35 broadens its host-range (18, 21, 22). It is likely that specific apoptotic suppressors borne by other baculoviruses are responsible for their abilities to infect specific insect species, tissues, or cells. Translation arrest is associated with abortive baculovirus infection in cell culture and has been observed in AcMNPV-infected BmN cells, AcMNPV hcf-1 deletion mutant-infected Tn368 cells, and Ld652Y cells infected with AcMNPV, HycuNPV, or BmNPV (44, 62, 68, 81). Translation arrest occurs as part of the innate immune response to virus infection in vertebrates, as a component of the interferon response signaled through Toll-like receptors (TLRs) (106), but it is unclear if it plays a similar role in insect defenses against virus infection. Recent advances in understanding insect immune responses to virus infection, that to date are limited to RNA viruses, and a few insect species and no Lepidoptera, suggests that the Toll-and Jak/Stat signaling pathways are involved in insect responses to virus infection (50, 136). The completion of the B. mori genome and identification of conserved genes involved in immune response in Drosophila, Aedes aegypti, and Anopheles gambiae may help to identify the genes responsible for lepidopteran immune responses to virus infection (15, 91, 129).