ABSTRACT
As a virus-encoded actin nucleation promoting factor (NPF), P78/83 induces actin polymerization to assist in Autographa californica multiple nucleopolyhedrovirus (AcMNPV) propagation. According to our previous study, although P78/83 actively undergoes ubiquitin-independent proteasomal degradation, AcMNPV encodes budded virus/occlusion derived virus (BV/ODV)-C42 (C42), which allows P78/83 to function as a stable NPF by inhibiting its degradation during viral infection. However, whether there are other viral proteins involved in regulating P78/83-induced actin polymerization has yet to be determined. In this study, we found that Ac102, an essential viral gene product previously reported to play a key role in mediating the nuclear accumulation of actin during AcMNPV infection, is a novel regulator of P78/83-induced actin polymerization. By characterizing an ac102 knockout bacmid, we demonstrated that Ac102 participates in regulating nuclear actin polymerization as well as the morphogenesis and distribution of capsid structures in the nucleus. These regulatory effects are heavily dependent on an interaction between Ac102 and C42. Further investigation revealed that Ac102 binds to C42 to suppress K48-linked ubiquitination of C42, which decreases C42 proteasomal degradation and consequently allows P78/83 to function as a stable NPF to induce actin polymerization. Thus, Ac102 and C42 form a regulatory cascade to control viral NPF activity, representing a sophisticated mechanism for AcMNPV to orchestrate actin polymerization in both a ubiquitin-dependent and ubiquitin-independent manner.
IMPORTANCE Actin is one of the most functionally important proteins in eukaryotic cells. Morphologically, actin can be found in two forms: a monomeric form called globular actin (G-actin) and a polymeric form called filamentous actin (F-actin). G-actin can polymerize to form F-actin, and nucleation promoting factor (NPF) is the initiator of this process. Many viral pathogens harness the host actin polymerization machinery to assist in virus propagation. Autographa californica multiple nucleopolyhedrovirus (AcMNPV) induces actin polymerization in host cells. P78/83, a viral NPF, is responsible for this process. Previously, we identified that BV/ODV-C42 (C42) binds to P78/83 and protects it from degradation. In this report, we determined that another viral protein, Ac102, is involved in modulating C42 ubiquitination and, consequently, ensures P78/83 activity as an NPF to initiate actin polymerization. This regulatory cascade represents a novel mechanism by which a virus can harness the cellular actin cytoskeleton to assist in viral propagation.
INTRODUCTION
Actin is an evolutionarily conserved protein found in all eukaryotic cells. It is present in both a monomeric form called globular actin (G-actin) and a polymeric form called filamentous actin (F-actin). G-actin can nucleate and polymerize to form F-actin with the help of nucleating factors, and the actin-related protein 2/3 complex (Arp2/3) is one of the best known actin nucleators (reviewed in reference 1). By itself, Arp2/3 is inefficient in nucleating G-actin. Nucleation promoting factors (NPFs), usually containing a regulatory domain at the N terminus and a catalytic domain (i.e., a WCA domain) at the C terminus, are the primary activators of Arp2/3 (reviewed in reference 2).
Pathogens frequently harness the host actin polymerization machinery to facilitate their propagation (reviewed in references 3 and 4). Baculoviridae is a family of double-stranded DNA viruses which has been divided into four genera, i.e., Alpha-, Beta-, Gamma-, and Deltabaculovirus. As an efficient eukaryotic expression vector and a widely used nontoxic biopesticide, Autographa californica multiple nucleopolyhedrovirus (AcMNPV) is the most studied Alphabaculovirus. A unique feature of AcMNPV infection is the virus-induced nuclear actin polymerization, which plays a key role in promoting progeny nucleocapsid morphogenesis (5, 6). P78/83, a viral capsid protein containing a WCA domain at the C terminus, is the NPF responsible for AcMNPV-induced actin polymerization (reviewed in reference 7).
In addition to the C-terminal WCA domain, P78/83 also contains an N-terminal degron motif that restricts its NPF activity by mediating robust ubiquitin-independent proteasomal degradation (8). Budded virus/occlusion derived virus (BV/ODV)-C42 (C42), a viral capsid protein, is a direct regulator of P78/83. By binding to the P78/83 degron motif, C42 protects P78/83 from proteasomal degradation, thus allowing P78/83 to function as a stable NPF and activate Arp2/3 (8).
Actin polymerization factors, including NPF, Arp2/3, and actin, are predominantly located in the cytoplasm by themselves. AcMNPV-induced nuclear actin polymerization requires these factors to accumulate in the nucleus during virus infection. Our previous findings revealed that P78/83 and Arp2/3 are redistributed to the nucleus by the viral late gene products C42 and Ac34 (9–11), respectively. Unlike P78/83 and Arp2/3, nuclear accumulation of actin is reportedly mediated by six viral early gene products, including IE-1, Pe38, He65, Ac004, Ac102, and Ac152 (12). Among them, Ac102, a 13-kDa viral early protein, is thought to play an essential role as deletion of ac102 from the AcMNPV genome resulted in the failure to translocate actin into the nucleus (13).
In this study, we determined that Ac102 is a novel regulator of actin polymerization during AcMNPV infection. Knockout of ac102 from the AcMNPV genome resulted in nuclear actin polymerization deficiency, as well as abnormal morphogenesis and distribution of capsid structures in the nucleus. These phenotypical changes are heavily dependent on the Ac102-C42 interaction but are not correlated with the nuclear accumulation of actin. Further investigation indicated that Ac102 suppresses K48-linked ubiquitination of C42 and consequently potentiates P78/83 availability as an NPF to induce actin polymerization.
RESULTS
Late viral gene products are required for AcMNPV-induced nuclear accumulation of actin.Previously, Ohkawa et al. reported that six early viral gene products translocate transiently expressed Bombyx mori actin into the nucleus of Sf21 cells during AcMNPV infection (12). To validate this conclusion, we investigated the spatial distribution of Spodoptera frugiperda actin (GenBank accession no. KY231202) in AcMNPV-infected Sf9 cells.
Green fluorescent protein (GFP)-tagged S. frugiperda actin (GFP-sf-actin) was transiently expressed in Sf9 cells by plasmid transfection. At 12 h posttransfection (hpt), the cells were either infected with AcMNPV or mock infected. To inhibit late viral gene expression, aphidicolin (APH) (14), an inhibitor of DNA synthesis, was added to the culture medium immediately after virus infection. At 24 h postinfection (hpi), the cells were fixed and subjected to confocal microscopy (CM). In mock-infected cells, GFP-sf-actin showed a predominantly cytoplasmic distribution pattern (Fig. 1A), presenting a significant contrast to the AcMNPV-infected cells, in which the nuclei were apparently enlarged, and GFP-sf-actin was evenly distributed in both the cytoplasm and nucleus (Fig. 1A). Surprisingly, when the infected cells were treated with APH, enlarged nuclei were not observed, and GFP-sf-actin was restricted to the cytoplasm (Fig. 1A). This phenotype suggested that late viral gene products are required for the nuclear accumulation of S. frugiperda actin during AcMNPV infection.
Late viral gene products are required for AcMNPV-induced nuclear accumulation of actin. (A) Impact of APH on nuclear accumulation of actin during AcMNPV infection. GFP-sf-actin was transiently expressed in Sf9 cells. At 12 hpt, the cells were either infected with AcMNPV (MOI of 5) or mock infected. After a 1-h incubation, the viral inoculum was removed, and DMSO or APH was added to the medium. At 36 hpi, the cells were fixed and subjected to CM. (B and C) Spatial distribution of actin during AcMNPV infection. GFP-sf-actin was transiently expressed in Sf9 cells. At 12 hpt, the cells were either infected with AcMNPV (MOI of 5) or mock infected. The spatial distribution of GFP-sf-actin was observed at the indicated time points (B), and the percentages of cells with accumulated GFP-sf-actin in the nucleus are presented in panel C. The bars represent the means and standard errors of the means for three independent experiments. Each experiment involved the quantification of at least 30 positively transfected cells. Scale bar, 7 μm.
To further validate our result, we performed a time-lapse study to determine the onset time of nuclear accumulation of GFP-sf-actin during AcMNPV infection. GFP-sf-actin was transiently expressed in Sf9 cells 12 h prior to AcMNPV infection. CM showed that GFP-sf-actin localized to the nucleus at 12 to 24 hpi (Fig. 1B and C), which is in accordance with the time point described by Goley et al., who showed that nuclear accumulation of enhanced GFP (EGFP)-actin began at 10 to 20 hpi in AcMNPV-infected TN-368 cells (6). As viral propagation proceeds into the late phase as early as 6 hpi (15), both findings support the conclusion that nuclear accumulation of actin occurs in the late infection phase, further indicating that late viral gene products are required for the nuclear accumulation of actin.
Ac102 is dispensable for AcMNPV-induced nuclear accumulation of actin.Gandhi et al. reported that Ac102 is the key factor involved in nuclear accumulation of actin (B. mori) during AcMNPV infection, as deletion of ac102 from the viral genome resulted in the failure to translocate actin into the nucleus of TN-368 cells (13).
To validate the role of Ac102 in the nuclear accumulation of actin in Sf9 cells, an ac102 knockout (ko) bacmid (vAcac102ko) (Fig. 2A) was constructed using the lambda red recombination system. PCR analysis with the indicated primer sets was performed to verify the replacement of ac102 with a chloramphenicol acetyltransferase (cat) expression cassette (Fig. 2A and B).
Generation and characterization of the ac102 knockout bacmid. (A). Diagram of ac102 knockout. The coding region (bp 137 to 211) of ac102 in the AcMNPV bacmid bMON14272 was replaced with a cat expression cassette. The knockout bacmid was termed vAcac102ko. The arrows indicate the locations of the PCR primers used to verify ac102 knockout. (B) Verifying the ac102 knockout. Two vAcac102ko candidates and bMON14272 were prepared and used as PCR templates to amplify DNA fragments with the indicated primer sets. (C) Diagram of transposed bacmids. Full-length and truncated ac102 or D61A controlled by the native ac102 promoter were transposed into vAcac102ko. To easily detect bacmid-harboring cells, GFP or mCherry expression cassettes controlled by the p10 promoter were simultaneously transposed into vAcac102ko. (D) Impact of ac102 knockout on ac101 (c42) transcription. Sf9 cells were transfected with the indicated bacmids. At 12, 18, and 24 hpt, the cells were subjected to qRT-PCR. Values are displayed as averages of ac101 transcripts compared to cellular 28s rRNA transcript levels (endogenous reference) from three independent experiments, with error bars indicating standard deviations. (E) Virus infectivity assay. The indicated bacmids were transfected into Sf9 cells. At 144 hpt, the cells were subjected to fluorescence microscopy. The viral supernatants were then collected and added to uninfected cells to initiate infection. At 72 hpi, the cells were subjected to fluorescence microscopy. (F) Virus growth kinetics. Viral stocks of vAcgfp and vAcac102res-gfp (MOI of 5) were used to infect Sf9 cells. At the indicated time points, viral supernatants were collected, and virus titers were calculated by an endpoint dilution assay. The data points indicate the averages of triplicate infections, and the error bars represent standard deviations. TCID50, 50% tissue culture infective dose.
A series of repaired bacmids was then constructed by transposing full-length or truncated Ac102 expression cassettes (controlled by the native ac102 promoter) into vAcac102ko (Fig. 2C). For easy detection of bacmid-harboring cells, a GFP or mCherry (mc) expression cassette (controlled by the p10 promoter) was transposed into vAcac102ko (yielding vAcac102ko-gfp or vAcac102ko-mc, respectively) and other repaired bacmids (Fig. 2C). To exclude the possibility that ac102 knockout disrupted the cis regulatory elements of neighboring genes, quantitative reverse transcription-PCR (qRT-PCR) was performed to compare the transcripts of ac101, which encodes C42 and is located downstream of ac102, in cells harboring vAcac102ko-gfp, a rescued bacmid with wild-type ac102 (vAcac102res-gfp), or a control bacmid (vAcgfp). At all the tested time points, no statistically significant differences were observed in ac101 transcripts among cells harboring the three different bacmids (Fig. 2D), indicating that ac102 disruption had no impact on the cis regulatory elements of ac101.
Transfection of vAcac102ko-gfp into Sf9 cells resulted in isolated EGFP-positive cells at 144 hpt (Fig. 2E), suggesting that vAcac102ko-gfp failed to induce transmissible infection. Adding viral supernatant to uninfected Sf9 cells also failed to produce any EGFP-positive cells at 72 hpi (Fig. 2E), confirming that vAcac102ko-gfp is propagation defective and that ac102 is an essential gene for AcMNPV propagation. As a control, vAcgfp generated high levels of EGFP-positive cells by either bacmid transfection or viral supernatant infection (Fig. 2E). vAcac102res-gfp behaved similarly to vAcgfp in a transfection-infection assay (Fig. 2E) and showed viral propagation kinetics similar to those of vAcgfp, as measured by a one-step growth curve assay (Fig. 2F), confirming that ac102 disruption has no off-target effects on other viral genes and is the only factor responsible for the loss of infectivity observed for vAcac102ko-gfp.
To assess the impact of Ac102 on actin localization, GFP-sf-actin was transiently expressed in Sf9 cells. At 12 hpt, the cells were transfected with vAcac102ko-mc, vAcac102res-mc, or vAcmc, all expressing mCherry (mc). At 48 hpt, the cells were fixed and subjected to CM. Unexpectedly, GFP-sf-actin localized to the nucleus in vAcac102ko-mc-harboring cells (evidenced by red fluorescence) and showed no apparent difference from the spatial distribution of GFP-sf-actin in vAcac102res-mc- or vAcmc-harboring cells (Fig. 3). This phenotype suggested that, unlike B. mori actin in TN-368 cells, the nuclear accumulation of S. frugiperda actin in AcMNPV-infected Sf9 cells is independent of Ac102.
Ac102 is dispensable for AcMNPV-induced nuclear accumulation of actin. Sf9 cells expressing GFP-sf-actin were transfected with vAcac102ko-mc, vAcac102res-mc, or vAcmc. At 48 hpt, the cells were fixed and subjected to CM. Scale bar, 7 μm.
Ac102 is indispensable for AcMNPV-induced nuclear actin polymerization.We next investigated whether Ac102 influences nuclear actin polymerization, which is a downstream event of nuclear actin accumulation.
In Sf9 cells transfected with vAcmc or vAcac102res-mc, phalloidin-stained F-actin was concentrated along the inner surface of the nuclear envelope (Fig. 4A). In contrast to the propagation-competent bacmids, vAcac102ko-mc failed to induce any visible nuclear F-actin (NF) in positively transfected cells. This phenotype suggested that although Ac102 is dispensable for nuclear accumulation of actin, it is indispensable for nuclear actin polymerization (Fig. 4A).
Ac102 is indispensable for AcMNPV-induced nuclear actin polymerization. (A) Ac102 is required for virus-induced nuclear actin polymerization. Sf9 cells expressing GFP-sf-actin were transfected with the indicated bacmids with an mCherry tag. At 48 hpt, the cells were fixed and stained with Alexa Fluor 633-phalloidin. Note that the upper row shows bacmids capable of inducing nuclear F-actin (NF), and the lower row shows bacmids that failed to induce NF. Arrowheads indicate bacmid-harboring cells. Scale bar, 7 μm. (B) Infectivity assay of bacmids harboring Ac102 truncation or mutation constructs. The indicated bacmids were transfected into Sf9 cells. At 144 hpt, bacmid-transfected cells were subjected to fluorescence microscopy, and viral supernatants were collected and titrated. The resulting viruses were used to infect new Sf9 cells. At 72 hpi, virus-infected cells were subjected to fluorescence microscopy.
To identify which specific sequence of Ac102 is required for nuclear actin polymerization, a series of repaired bacmids encoding truncated Ac102 was transfected into GFP-sf-actin-expressing cells. At 48 hpt, the cells were fixed and stained with phalloidin. CM showed that nuclear GFP-sf-actin was present in all the bacmid-transfected cells, while NF was restricted to the cells bearing vAcac102Δ1–10-mc (a bacmid encoding Ac102 with a deletion of amino acids [aa] 1 to 10), vAcac102Δ11–20-mc, vAcac102Δ21–30-mc, vAcac102Δ41–50-mc, vAcac102Δ108–122-mc, vAcac102Δ113–122-mc, and vAcac102Δ118–122-mc (Fig. 4A and Table 1), indicating that aa 31 to 40 and 51 to 107 of Ac102 are required for nuclear actin polymerization. Accordingly, these two fragments were also essential for viral infectivity (Fig. 4B and Table 1). Characterization of the propagation-competent bacmids revealed that their NF percentages varied. NF was present in approximately 50% of the vAcac102res-mc- or vAcmc-harboring cells, as well as in 48.6% of the vAcac102Δ11–20-mc-harboring cells and 42.9% of the vAcac102Δ21–30-mc harboring cells (Table 1). Endpoint dilution assays demonstrated that the four bacmids generated similar viral titers at 96 hpi (Fig. 4B and Table 1). In vAcac102Δ1–10-mc-, vAcac102Δ108–122-mc-, and vAcac102Δ113–122-mc-harboring cells, the NF percentage decreased to 9.1%, 5.9%, and 12.5%, respectively, which was in striking contrast to the high percentage observed in vAcac102Δ41–50-mc-positive (76.9%) and vAcac102Δ118–122-mc-positive (74.6%) cells (Table 1). Interestingly, both the low- and high-NF-percentage bacmids generated low viral titers compared with the moderate-NF-percentage bacmids (Fig. 4B and Table 1).
Infectivity and actin dynamics of Ac102 truncations and mutantsa
We also identified a lethal Ac102 mutation, vAcac102D61A (D61 is one of the most conserved amino acids among all sequenced Ac102 homologs) (see Fig. S1 in the supplemental material), which produced no NF (Fig. 4A and Table 1) and completely abolished virus infectivity (Fig. 4B and Table 1).
Ac102 modulates the morphogenesis and spatial distribution of viral capsid structures.Given that Ac102 is indispensable for virus-induced nuclear actin polymerization, which assists in viral nucleocapsid morphogenesis, we next investigated the influence of Ac102 on the ultrastructure of the nucleocapsid. Electron microscopy (EM) analysis of vAcmc- and vAcac102res-mc-harboring cells showed that rod-shaped, electron-dense nucleocapsids were preoccluded or occluded into virions at the ring zone, presenting a dramatic contrast to the fiber-shaped, electron-lucent capsid structures packed into bundles and restricted to the nuclear periphery in vAcac102ko-mc-, vAcac102D61A-mc-, and vAcac102Δ31–40-mc-harboring cells (Fig. 5A; see also Fig. S2). An examination of cells harboring vAcac102Δ41–50-mc revealed that unbundled, rod-shaped capsid structures with frequent electron-dense content were dispersed in the virogenic stroma, whereas these structures were occluded into virions at the ring zone in vAcac102Δ118–122-mc-harboring cells (Fig. 5A), thus exhibiting no apparent difference from the ultrastructures observed in vAcmc- and vAcac102res-mc-harboring cells.
The impact of Ac102 on viral ultrastructure and capsid protein distribution in the nucleus. (A) Ultrastructure of cells harboring the indicated bacmids. Sf9 cells were transfected with the indicated bacmids. At 120 hpt, the cells were subjected to EM. The boxed region in the left frame of each pair of frames is displayed magnified in the right frame. Arrows indicate normal-appearing nucleocapsids or capsid structures with or without electron-dense content. RZ, ring zone; VS, virogenic stroma; NM, nuclear membrane. (B) Viral capsid protein distribution in the nucleus. Plasmids encoding viral capsid proteins were cotransfected with the indicated bacmids. At 48 hpt, the cells were subjected to an immunofluorescence assay using anti-HA. White arrowheads indicate double-transfected cells (bacmids and plasmids encoding viral capsid proteins). Scale bar, 7 μm.
To investigate the impact of Ac102 on the distribution of viral capsid proteins in the nucleus, plasmids encoding hemagglutinin (HA)-tagged VP39, PCNA, VP1054, VLF-1, 38K, VP80, P24, and EC27 were cotransfected with vAcmc (a positive control bacmid), vAcac102ko-mc, or vAcac102D61A-mc. CM showed that in all the bacmid-transfected cells, all the tested minor capsid proteins were dispersed throughout the nucleus (Fig. 5B, frames b to h). Only the major capsid protein VP39 was restricted to the nuclear periphery in vAcac102ko-mc- and vAcac102D61A-mc-harboring cells, while it was dispersed throughout the nucleus in vAcmc-harboring cells (Fig. 5B, frame a).
Taken together, our results suggested that Ac102 not only participates in nucleocapsid morphogenesis but also modulates the distribution of capsid structures in the nucleus. In fact, the morphogenesis and spatial distribution of capsid structures in vAcac102ko-mc- and vAcac102D61A-mc-bearing cells strongly resembled the ultrastructure previously seen in cells bearing vAcc42ko or vAc54KO (ac54 encodes VP1054, which interacts with C42) (16, 17). This resemblance prompted us to hypothesize that Ac102 is functionally connected to C42.
Ac102 and C42 are present in the same protein complex.To explore whether Ac102 interacts with C42, GFP-Ac102-V5 was coexpressed with C42-FLAG in Sf9 cells. Coimmunoprecipitation (co-IP) using an anti-FLAG antibody showed that Ac102 immunoprecipitated with C42 in both the presence and absence of viral infection (Fig. 6A), suggesting that Ac102 and C42 are present in the same protein complex, although it is unclear whether this is through a direct or indirect interaction. To identify which specific sequence of Ac102 is required for the interaction with C42, a series of Ac102 truncation constructs was coexpressed with C42. A co-IP assay demonstrated that deletion of aa 51 to 117 from Ac102 failed to immunoprecipitate C42 (Fig. 6B), while the region of aa 51 to 117 itself could immunoprecipitate C42 (Fig. 6C), suggesting that aa 51 to 117 are required for Ac102 interaction with C42. Coincidently, the effect of Ac102 on nuclear actin polymerization is also associated with aa 51 to 117 as the absence of aa 51 to 103 completely abolished NF, and a deletion of aa 103 to 117 severely decreased NF (Fig. 4A and Table 1). This coincidence suggested that Ac102-mediated modulation of nuclear actin polymerization is heavily, although not entirely, dependent on the Ac102-C42 interaction. In support of this hypothesis, Ac102D61A severely compromised the interaction between Ac102 and C42 (Fig. 6D) and failed to rescue NF formation when it was introduced to vAcac102ko (Fig. 4A).
Characterization of the Ac102-C42 interaction. (A) C42-FLAG and GFP-Ac102-V5 were coexpressed in Sf9 cells. At 24 hpt, the cells were either mock infected or infected with vAcmc stock. At 48 hpt, the cells were harvested and subjected to co-IP using anti-FLAG. The immunoprecipitated proteins were probed with the indicated antibodies. (B to D) C42-FLAG was coexpressed with the indicated Ac102 truncation or mutation constructs in Sf9 cells. At 48 hpt, the cells were harvested and subjected to co-IP using anti-V5 (B and C) or anti-FLAG (D). The immunoprecipitated proteins were probed with the indicated antibodies. WCL, whole-cell lysate.
Ac102 modulates C42 proteasomal degradation by suppressing K48-linked ubiquitination of C42.To investigate how Ac102 interferes with C42, C42 was transiently expressed in Sf9 cells in the presence of Ac102 at different concentrations. Western blot analysis showed that Ac102 could increase C42 protein abundance in a dose-dependent manner (Fig. 7A). Since the ubiquitin (Ub)-proteasome pathway (UPP) is the primary mechanism for protein degradation, we next investigated whether Ac102 modulates C42 protein abundance by interfering with C42 ubiquitination.
Characterization of Ac102 ubiquitination. (A) Impact of Ac102 on C42 protein abundance. C42-FLAG was coexpressed with increasing amounts of GFP-Ac102-V5 (0, 0.2, 0.6, or 1.2 μg of plasmid). At 48 hpt, the cells were subjected to Western blot analysis and probed with the indicated antibodies. (B to F) Characterization of Ac102 ubiquitination. Plasmids (B, D, E, and F) or a bacmid (C) encoding the indicated proteins was cotransfected into Sf9 cells. At 44 hpt, DMSO or PS-341 was added to the medium. After 4 h of incubation, the cells were harvested and subjected to co-IP. The immunoprecipitated proteins were probed with the indicated antibodies. In panel E, to quantify the impact of Ac102 on C42 ubiquitination, densitometry analysis was performed by using ImageJ software (https://imagej.nih.gov/ij/). The density of the smeared bands (ubiquitinated C42) was normalized to the that of the immunoprecipitated C42-FLAG bands. The bars represent the means and standard errors of the means for three independent experiments. *, P < 0.05. (G) Sf9 cells were transfected with the indicated bacmids. At 48 hpt, the cells were harvested and subjected to Western blot analysis using the indicated antibodies.
To detect C42 ubiquitination, C42-FLAG was coexpressed with Myc-tagged ubiquitin (Myc-Ub) (8). In contrast to results with a dimethyl sulfoxide (DMSO) treatment, immunoprecipitated C42-FLAG appeared as a smeared band in the presence of PS-341 (Fig. 7B), a proteasomal degradation inhibitor (8). This observation indicated that C42 could be ubiquitinated. To validate C42 ubiquitination in virus-infected cells, an HA-tagged C42 expression cassette controlled by the native ac101 promoter was transposed into bMON14272 (Invitrogen), a commercially available bacmid harboring the wild-type AcMNPV genome. The resulting virus (vAcc42HA) was used to infect Sf9 cells in the presence or absence of PS-341. Co-IP using an anti-HA antibody showed that C42-HA appeared as a smeared band in the presence of PS-341 when probed with an anti-ubiquitin antibody (Fig. 7C), confirming that C42 is ubiquitinated in virus-infected cells. To further identify which linkage types are present on ubiquitinated C42, C42 was coexpressed with Myc-Ub-K48 (a ubiquitin mutant bearing arginine substitutions on all of its lysine residues except the one at position 48 and thus expected to maintain the K48-linked, while completely abolishing the K63-linked, polyubiquitination of proteins) or Myc-Ub-K63 (a ubiquitin mutant bearing arginine substitutions on all of its lysine residues except the one at position 63 and thus expected to maintain the K63-linked, while completely abolishing the K48-linked, polyubiquitination of proteins) (18). A co-IP assay showed that both Ub-K48 (Fig. 7D, lane 3) and Ub-K63 (Fig. 7D, lane 4) could ubiquitinate C42, indicating that C42 undergoes both types of ubiquitination.
To assess the impact of Ac102 on C42 ubiquitination, C42 was coexpressed with Myc-Ub in the presence or absence of Ac102. A comparison of the immunoprecipitated C42 under the two conditions revealed that Ac102 apparently decreased the C42 ubiquitination level (Fig. 7E). Since C42 underwent both K48- and K63-linked ubiquitination, we next investigated which linkage types Ac102 inhibits. Co-IP showed that in the presence of Ac102, levels of both K48- and K63-linked C42 ubiquitination were decreased (Fig. 7F). Since K48-linked ubiquitination is the primary linkage type for UPP-mediated protein degradation (reviewed in reference 19), Ac102 potentiates C42 availability by suppressing its UPP-mediated degradation. Notably, Ac102-mediated suppression of K48-linked C42 ubiquitination is dependent on the Ac102-C42 interaction as Ac102D61A showed no apparent impact on K48-linked ubiquitination of C42 (Fig. 7F). In contrast, Ac102D61A effectively decreased K63-linked ubiquitination of C42 (Fig. 7F), suggesting that, unlike K48-linked ubiquitination, Ac102 inhibition of K63-linked ubiquitination of C42 is independent of the Ac102-C42 interaction.
To validate the role of Ac102 in modulating C42 in virus-infected cells, the aforementioned C42 expression cassette was transposed into vAcac102ko and vAcgp64ko (a gp64-null bacmid, kindly provided by Zhihong Hu of Wuhan Institute of Virology, CAS) (20), and the resulting vAcac102ko-C42HA and vAcgp64ko-C42HA bacmids were transfected into Sf9 cells. At 24 hpt, the cells were harvested and subjected to Western blot analysis. Compared with the level in vAcgp64ko-C42HA-transfected cells, HA-tagged C42 decreased sharply in cells transfected with vAcac102ko-C42HA (Fig. 7G), suggesting that Ac102 potentiated C42 availability in virus-infected cells.
Accordingly, P78/83 protein levels decreased simultaneously with C42 in vAcac102ko-C42HA-transfected cells (Fig. 7G), which is likely because C42 directly protects P78/83 from proteasomal degradation in an ubiquitination-independent manner, as we previously reported (8).
DISCUSSION
Nuclear actin polymerization is a unique feature of AcMNPV infection, and an ever-expanding list of viral proteins has been identified to be involved in this process. In this study, we identified that Ac102 is a novel factor involved in nuclear actin polymerization by forming a regulatory cascade with C42 to control P78/83 availability as an NPF.
To induce nuclear actin polymerization, AcMNPV employs different approaches to translocate cytoplasmic actin polymerization factors to the nucleus. Nuclear accumulation of actin has been reported to be attributed to early viral gene products, and Ac102 is essential for this process (12, 13). However, our study demonstrated that late viral gene products are also involved and that Ac102 is dispensable for nuclear actin accumulation. This discrepancy could be attributed to at least two factors. The first is that different actin clones were used: B. mori actin versus S. frugiperda actin (amino acid identity of 97%). The second is that the transiently expressed actin was expressed at different times: Ohkawa et al. transiently expressed actin at 2 hpi (12), whereas we expressed actin at 12 h prior to virus infection. Because actin is a housekeeping protein and resides in cells irrespective of virus infection, we think that transiently expressing actin prior to viral infection better simulates the natural process of virus infection of cells.
Although NF benefits AcMNPV propagation by assisting in nucleocapsid morphogenesis (5, 6), our data indicate that a high NF level could also compromise the virus titer (Fig. 4B and Table 1). Bacmids that induce NF overload, in particular vAcac102Δ118–122-mc, showed indistinguishable levels of nucleocapsid morphogenesis compared with vAcmc or vAcac102res-mc, indicating that their compromised virus titers are due to other defects in the viral propagation cycle.
Nuclear G-actin is required for the transcriptional activity of RNA polymerase II (21), whereas the polymerization of G-actin to F-actin inhibits transcription by RNA polymerase II (22). AcMNPV employs host RNA polymerase II to transcribe immediate early viral genes (e.g., ie-1 and ie-2), and the resulting viral proteins serve as trans-activators to promote later viral gene expression in a cascade fashion (23). Beperet et al. reported that AcMNPV superinfection exclusion occurs at 16 to 20 hpi of the first virus, concurrent with the onset of nuclear actin polymerization, and that inhibition of nuclear actin polymerization by cytochalasin D could rescue AcMNPV superinfection (24). They proposed that nuclear actin polymerization induced by the first virus could sequester nuclear G-actin and consequently prevent host RNA polymerase II from transcribing the immediate early genes of the second virus, thus resulting in superinfection exclusion (24). Based on the aforementioned theory, excessive NF induced by vAcac102Δ41–50-mc or vAcac102Δ118–122-mc could overdeplete nuclear G-actin, resulting in severely reduced transcription of immediate early viral genes of descendant virus, thus compromising the total virus titer. Taken together, current evidence suggests that NF has dual effects on virus propagation: on one side, NF benefits parental virus propagation by assisting in nucleocapsid morphogenesis; on the other side, NF restricts descendant virus propagation (as well as superinfection) by suppressing the viral transcription cascade.
Ubiquitination is a multifunctional biological process that occurs in eukaryotic cells. K48-linked ubiquitination is restricted to mediate protein degradation by the proteasome (reviewed in reference 19), while K63-linked ubiquitination may regulate a variety of processes, such as protein trafficking, gene transcription, and viral infection (reviewed in reference 25). Wiskott-Aldrich syndrome protein homologue (WASH), a mammalian NPF involved in endosome-to-Golgi transport, is locked in an autoinhibitory conformation under steady-state conditions. K63-linked ubiquitination of WASH disrupts the autoinhibitory conformation and releases its WCA domain to activate Arp2/3 for actin polymerization (26). In this study, we showed that Ac102 participates in modulating actin polymerization by suppressing K48-linked ubiquitination of C42, thus forming a regulatory cascade with C42 to control P78/83 availability as an NPF. Interestingly, the results of our study indicated that in addition to K48-linked ubiquitination, C42 also undergoes K63-linked ubiquitination, prompting us to consider that C42 may regulate P78/83-induced actin polymerization in a yet to be determined, K63-linked ubiquitination-dependent manner (Fig. 8). Our data showed that bacmids encoding Ac102 mutants with intact aa 51 to 117 (vAcac102Δ1–10-mc, vAcac102Δ31–40-mc, vAcac102Δ41–50-mc, and vAcac102Δ118–122-mc), the sequence required for Ac102-C42 interaction, exhibited abnormal NF levels compared with the level of the wild-type virus (Fig. 4A and Table 1). Such phenotypes could possibly be attributed to encoded Ac102 mutants in the aforementioned bacmids that modulate K63-linked ubiquitination of C42 independent of Ac102-C42 interaction and consequently influence P78/83-induced actin polymerization.
Model of a regulatory cascade that controls actin polymerization during AcMNPV infection. Ac102 sits at the top of a regulatory cascade by inhibiting both K48- and K63-linked ubiquitination of C42. K48-linked ubiquitination promotes the degradation of C42 by UPP, decreasing C42 availability. The reduced C42 levels then compromises its ability to protect P78/83 from ubiquitination-independent proteasomal degradation and eventually results in the attenuation of P78/83-induced actin polymerization. Unlike K48-linked ubiquitination, which is primarily associated with UPP-mediated protein degradation, K63-linked ubiquitination of C42 is hypothesized to modulate P78/83-induced actin polymerization via a yet unknown mechanism, as indicated by the question mark.
In summary, our study revealed a novel regulatory cascade consisting of Ac102 and C42 (as summarized in Fig. 8), which extends our understanding of how baculovirus manipulates the host actin polymerization machinery to maximize its value to viral propagation.
MATERIALS AND METHODS
Cell culture, transfection, and virus infection.Sf9 cells were maintained at 27°C in Grace's medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Invitrogen). Plasmids or bacmids were transfected into Sf9 cells using Cellfectin II (Invitrogen) according to the manufacturer's protocols. Viral supernatants or stocks (multiplicity of infection [MOI] of 5 unless otherwise indicated) were incubated with uninfected Sf9 cells at 4°C for 60 min. After viral attachment, the cells were replenished with complete medium containing dimethyl sulfoxide (DMSO) or APH (5 μg/ml). Virus titration and infectivity assays were performed as described previously (9).
Preparation of the ac102-knockout bacmid.To delete ac102 from AcMNPV, the lambda red recombination system was employed as described previously (9). The recombinant bacmid was verified by PCR.
Construction of plasmids and recombinant bacmids.A standard molecular cloning protocol was used to prepare plasmids encoding the indicated proteins. To prepare recombinant bacmids, a Bac-to-Bac protocol was employed. In brief, gene expression cassettes were cloned into pFastBac-dual (Invitrogen), and the resulting shuttle vectors were transformed into DH10B Escherichia coli cells harboring bMON14272 or vAcac102ko to generate transposed bacmids.
qRT-PCR assay.trizol reagent (invitrogen) was used to extract total rna from bacmid-transfected sf9 cells according to the manufacturer's protocols. the resulting total rna samples (1 μg) were processed with a quantitect probe rt-pcr kit to quantify ac102 and 28s rrna (endogenous reference) transcripts. the primer sets used for qrt-pcr were ttcgttgaacagcaccagtt and aaacgtggtttcgagcagtt to quantify c42 and ctggcttgatccagatgttcag, ggatcgataggccgtgctt to quantify 28s rrna.
Electron microscopy assay.Bacmid-transfected Sf9 cells were harvested, fixed, dehydrated, embedded, sectioned, and stained as described by O'Reilly et al. (27). The samples were examined with a Tecnai G2 transmission electron microscope at an accelerating voltage of 80 kV.
Western blotting and co-IP assays.Cells were rinsed twice with phosphate-buffered saline (PBS) and lysed in radioimmunoprecipitation assay (RIPA) buffer (150 mM sodium chloride, 1% Triton X-100, 0.5% sodium deoxycholate, 50 mM Tris, pH 8.0) with cOmplete protease inhibitor cocktail (Roche). Total proteins were quantified by a Quick Start Bradford assay (Bio-Rad). Cell lysates containing 50 μg of total protein were mixed with 2× Laemmli buffer (Bio-Rad) and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The proteins were then transferred to nitrocellulose membranes (Millipore), blocked in 0.5% nonfat dry milk, and incubated overnight at 4°C with the following antibodies: anti-V5 (Invitrogen), anti-FLAG (M2; Sigma), anti-ubiquitin (Abcam), anti-EGFP and anti-actin (Santa Cruz), and anti-P78/83 (Abmart). After incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson Laboratory), the membranes were developed by enhanced chemiluminescence (Pierce). For co-IP assays, 2,000 μg of total protein was incubated with 2 μg of the antibodies indicated on Fig. 6 and 7 and 50 μl of protein G-agarose (Millipore) overnight at 4°C. The protein-binding agarose was then extensively washed with RIPA buffer, mixed with 2× Laemmli buffer, and subjected to Western blot analysis.
Immunofluorescence confocal microscopy and F-actin staining.Cells grown on coverslips were washed with PBS, fixed in 3.7% formaldehyde for 10 min at room temperature, and permeabilized with 0.1% Triton X-100 for 5 min. The cells were then blocked in 1% normal goat serum (Boster) in PBS for 30 min on ice.
For immunofluorescence assays, the cells were incubated with anti-HA primary antibodies (1:500 dilutions; Sigma). An Alexa Fluor 488-conjugated anti-mouse antibody (1:500 dilution; Invitrogen) was used as the secondary antibody. For F-actin staining, the cells were incubated with a 1/40 dilution of Alexa Fluor 633-phalloidin (Invitrogen) at room temperature inside a covered container for 20 min. To visualize DNA in the nucleus, the cells were incubated with Hoechst 33258 (1:1,000 dilution; Invitrogen) for 10 min in a dark room.
After being extensively washed with PBS, the samples were subjected to observation by CM using a PerkinElmer UltraVIEW VoX microscope.
ACKNOWLEDGMENTS
We are grateful to Yuntao Wu of George Mason University for suggestions on our study. We also thank Ding Gao, Anna Du, Bichao Xu, and Pei Zhang of the core facilities in the Wuhan Institute of Virology, CAS, for their technical assistance with CM and EM.
This work was supported by grants from the National Natural Science Foundation of China (31770170 and 31470261 to Y.W. and 31621061 to X.C.).
FOOTNOTES
- Received 16 February 2018.
- Accepted 23 March 2018.
- Accepted manuscript posted online 4 April 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/JVI.00005-18.
- Copyright © 2018 American Society for Microbiology.
