Aggregation of AcMNPV LEF-10 and Its Impact on Viral Late Gene Expression

The Autographa californica multiple nucleopolyhedrovirus (AcMNPV) late expression factor gene lef-10 has been identified to be required for viral late gene expression by transient expression assay. Our previous work has shown that the gene product LEF-10 can form very stable high-molecular-weight complexes, but the structure and function of the protein remain unknown. In this study, we demonstrated that LEF-10 was essential for the replication of AcMNPV, and its truncated fragment containing amino acid residues 1 to 48 were sufficient to support the virus survival. Our data also suggested that the LEF-10 could spontaneously aggregate to form punctate spots in virus infected Sf9 cells at low frequency, and the aggregation of the protein could be induced by LEF-10 over-expression. When the protein aggregated to form punctate spots, soluble LEF-10 proteins were depleted and this could result in the down-regulation of viral late gene expression.


Introduction
Baculoviruses are insect-specific viruses, mainly infecting Lepidopteran insects. Their gene expression can be divided into three phases: early, late and very late periods. Baculoviruses utilize host RNA polymerase II as well as virus-encoded RNA polymerase to transcribe viral genes during infection [1]. Host RNA polymerase II recognizes baculovirus early promoters, and the virus-encoded RNA polymerase recognizes the late and very late promoters [2,3]. The expression of viral late/very late genes requires the accumulation of viral early gene products including the virus-encoded RNA polymerase.
Lef-10 is thought to be one of the genes positively selected during baculovirus evolution [22], and its homologs are found in the genomes of all Group I and most Group II nucleopolyhedroviruses (NPV) and granuloviruses (GV) [3]. It has been reported that lef-10 of Bombyx mori nucleopolyhedrovirus (BmNPV) is essential for the viral DNA replication and gene expression [23,24].
Previously, we reported that LEF-10 expressed in Escherichia coli and Spodoptera frugiperda (Sf9) cells formed high-molecular-weight complexes [25]. The complexes were very resistant to the treatment of 6M Urea, or 2% SDS with 100 mM DTT (dithiothreitol). This phenomenon has never been reported for any other LEFs or other baculovirus proteins to the best of our knowledge.
To understand the function and characteristics of LEF-10, we constructed a lef-10 knockout AcMNPV bacmid, identified the regions involved in the function and aggregation of the protein, and investigated the correlation of LEF-10 aggregation with its function in late gene expression regulation.

Cells and viruses
Sf9 cells were maintained in SFX insect medium (Thermo Scientific HyClone) supplemented with 1% fetal bovine serum (FBS) at 27°C. Recombinant baculoviruses were produced by homologous recombination between co-transfected pTriEx-1.1 or pBAC-5 (Novagen) based plasmid and linearized bacmid [26] using FuGENE HD Transfection Reagent (Promega). Baculoviruses generated by homologous recombination were harvested at 96 hours post transfection. Multiplicities of infection (MOI) used in baculovirus infection were 0.05 for virus amplification and 3 for protein expression.
To perform immunofluorescence assay, the lef-10 promoter and coding region (-147 to 235) was cloned into pBAC-5 between the BglII and XhoI sites. The constructed plasmid was named as pBAC5/lef10p:lef10-His, which could produce His-tagged LEF-10 controlled by its intrinsic promoter.

Growth kinetic analysis of LEF-10 repaired viruses
To compare the one step growth curve of the wild type and LEF-10 repaired recombinant viruses, Sf9 cells were infected with the indicated virus at an MOI of 1. Cell culture media were harvested at various time points and analyzed for the presence of infectious virions by endpoint dilution. The averages of the infectious titers from three independent infections were calculated and plotted into graphs.

Fluorescence microscopy
Sf9 cells were infected with recombinant baculoviruses. The expression of fluorescent proteins was observed using fluorescence microscope (Leica DM5000B upright microscope or Olympus CKX41 inverted microscope).

Immunofluorescence assay
To investigate whether LEF-10 aggregation in vivo is caused by GFP fusion, a smaller tag, 8×His instead of GFP, was introduced into the c-terminus of LEF-10, and the protein expression was detected by immunofluorescence assay. Briefly, recombinant baculovirus vAc/lef10p:lef10-His was generated by homologous recombination using pBac5/lef10p:lef10-His and BacmidΔlef10. Sf9 cells (8×10 5 cells/dish) in 35 mm plate were infected with vAc/lef10p:lef10-His at an MOI of 2. After incubation for 24 hours at 27°C, cells were washed 3 times with phosphate-buffered saline (PBS: 50 mM Tris, 150 mM NaCl, pH 7.6). The cells were fixed by adding ice-cold 95% ethanol and incubating for 3 min. Cells were then permeabilized with 0.1% Triton X-100 in PBS at room temperature for 15 min. After washing 3 times with PBS, the cells were incubated overnight at 4°C with anti-His mAb at 1:1000 dilution in PBS. After washing, bound antibody was detected with a goat anti-mouse IgG conjugated to FITC (Beijing CoWin Biotech).
To visualize the cell nuclei, cells were then stained with 10 μg/ml of DAPI in the dark for 30 min. Stained cells were observed under fluorescence microscope.

Flow cytometry
Sf9 cells were harvested at 72 hours post infection (hpi) by centrifugation at 300 g for 2 min, and then re-suspended in PBS. Data collection was performed on a CyFlow Cube 6 flow cytometer (Partec) at an excitation wavelength of 488 nm. Cells expressing GFP fusion protein were detected in the FL1 channel at emission wavelength of 530 nm. Cells expressing mCherry were detected in the FL2 channel at emission wavelength of 585 nm. Data were collected from at least 2×10 4 cells for each sample and data analysis was performed offline using FCS Express 4.

Semi-quantitative RT-PCR
To investigate the effect of LEF-10 aggregation on late gene transcription, semi-quantitative RT-PCR was performed. Total RNAs were isolated with the TRIzol RNA extraction kit (Beijing CoWin Biotech) from 1×10 6 Sf9 cells infected by vAcΔlef10/actinp-p10p:lef10 1-65 -GFP or vAc-Δlef10/actinp-p10p:lef10-GFP (Fig 1D) at 48 hpi. The RNA samples were then treated with RNase-Free DNaseI (Thermo). Reverse transcription was performed using the 5× All-In-One RT MasterMix (ABM), employing 2 μg total RNA as the template per reaction. The target DNA fragments were then amplified by PCR using the according primer pairs (listed in Table 1). After 25-29 cycles of amplification, the PCR samples were analyzed by 1% agarose gel electrophoresis.

AcMNPV lef-10 is essential for virus replication
It has been reported that lef-10 of BmNPV is required for its viral replication [23,24]. To investigate the role of LEF-10 in AcMNPV infection, a lef-10 knockout AcMNPV bacmid (Bacmid-Δlef10) was generated using the ET-recombination system in E. coli (Fig 1A). Co-transfection of Sf9 cells with BacmidΔlef10 and pTriEx/GFP failed to produce infectious baculovirus (Fig  2A left), in agreement with the previous reports that lef-10 is essential for baculovirus replication [23].

Gene
Primer Name Primer Sequence doi:10.1371/journal.pone.0154835.t001 infected with the supernatant obtained from the transfected cell culture and observed at 72 hpi. Visible cytopathic effects (CPE) were observed under bright field (Fig 2A middle), demonstrating that the function of native lef-10 was compensated by lef10-GFP.
As the expression level of LEF10-GFP under the control of its intrinsic promoter was quite low and hardly detectable by fluorescence microscope (Fig 2A middle), this made it difficult to characterize the protein in further studies. To increase the expression level of LEF-10, we used pTriEx/lef10-GFP to introduce lef10-GFP driven by tandem chicken actin promoter (recognized by cellular RNA polymerase II) and p10 promoter into the BacmidΔlef10 (Fig 1D). Using this recombinant virus, the expression of LEF10-GFP could be initiated by the chicken actin promoter after virus infection, and the protein could also be over-produced under the control of the strong p10 promoter in late infection. As shown in Fig 2A (the right images), infectious viruses were successfully rescued, and the over-expressed LEF10-GFP was also easily observed using the fluorescence microscope.
To investigate the subcellular localization of LEF-10, His-tagged LEF-10 under the control of the lef-10 intrinsic promoter was introduced into BacmidΔlef10 to rescue the baculovirus, and the expression of His-tagged LEF-10 was detected by immunofluorescence assay using anti-His mAb and FITC conjugated secondary antibodies. Nuclear localization of the fluorescent protein were observed in most infected cells (Fig 2B), implying that LEF-10 may function in the nucleus.

Identification of the essential region for LEF-10 function
To identify the LEF-10 functional region for the virus replication, a set of truncated lef-10 gene fragments (lef-10 14-78 , lef-10 27-78 , lef-10 1-39 , lef-10 1-47 , lef-10 1-48 and lef-10 1-52 , illustrated in Fig 1F) were inserted into pTriEx/GFP between NcoI and BamHI sites, and then introduced into the polh locus to repair BacmidΔlef10 by homologous recombination. Sf9 cells were infected with supernatants obtained from the transfected cells and observed by fluorescence microscope at 72 hpi. Visible CPE and green fluorescence were observed in the cells infected with the recombinant viruses expressing lef-10 1-48 and lef-10 1-52 , but not in the cells infected with the recombinant viruses containing lef-10 14-78 , lef-10 27-78 , lef-10 1-39 and lef-10 1-47 ( Fig  2C). For those samples gave negative results, the supernatants obtained from the transfected cells were used to infect Sf9 cells for 3 passages, and still no CPE were observed. These results demonstrated that amino acids 1 to 48 were sufficient to recover the function of LEF-10 and the C-terminus (aa 49-78) of LEF-10 was not essential for the production of infectious viruses.
Growth kinetic analysis of LEF-10 repaired viruses confirmed that knockout of lef-10 was lethal for AcMNPV, and the function of LEF-10 could be repaired by LEF10-GFP. Although over-expressing of LEF10-GFP and LEF10 1-48 -GFP could rescue the recombinant viruses, the virus titers decreased dramatically (Fig 2D). The replication level of the virus expressing truncated LEF10 1-48 -GFP reduced two orders of magnitude compared to LEF10-GFP, indicating that the C-terminus (aa 49-78) of LEF-10 was not essential for the producing infectious viruses but was beneficial for the virus production at high levels.

Aggregation of LEF-10 in baculovirus infected cells
To investigate the aggregation of LEF-10 in insect cells, the natural host of baculoviruses, the recombinant baculovirus expressing LEF10-GFP driven by its intrinsic lef-10 promoter ( Fig  1C) was used to infect Sf9 cells. LEF10-GFP was observed to aggregate in punctate dots in some infected cells (less than 1%), although the protein was uniformly distributed in the majority of the infected cells at 24 hpi (Fig 3A). Immunofluorescence assay using anti-His mAb showed that LEF10-His, the protein fused with a much smaller tag than GFP, also formed punctate dots similar to the LEF10-GFP fusion protein (Fig 2B). The results here suggested that the aggregation of LEF-10 could be a natural process in baculovirus infected cells. As no strong DAPI signals were found colocalized with the punctate dots (Fig 2B), it indicated that LEF-10 complexes were not stainable with DAPI and the aggregation of LEF-10 was not associated with apoptotic bodies which could easily be stained by DAPI.
We also observed that over-expressed LEF-10 driven by the p10 promoter aggregated more frequently (about a quarter to a half of the cells) than the LEF-10 expressed under the control of its intrinsic promoter (Fig 3B). In comparison, the truncated protein LEF-10 1-65 failed to form punctate dots in Sf9 cells even when it was over-produced under the control of p10 promoter (Fig 3C). Deletion of the C-terminal region disrupted the aggregation of LEF-10 in infected cells (LEF-10 1-65 in Fig 3C), but did not interrupt the protein's function in virus replication (the replication level of the virus expressing LEF-10 1-65 detected by virus titration was comparable to the virus expressing full-length LEF-10, data not shown), suggesting that aggregation of LEF-10 was not essential for the viral replication.

Aggregation of LEF-10 down-regulated its function
To explore if the aggregation of LEF-10 was related to its function, two recombinant baculoviruses, vAcΔlef10/lef10p:lef10-GFP/p10p:mCherry and vAcΔlef10/lef10p:lef10 1-65 -GFP/p10p: mCherry, were constructed (Fig 1E). In these recombinant viruses, expression of LEF-10 or its truncated form was controlled by its intrinsic promoter, and mCherry driven by the p10 promoter was used as a reporter to reflect the very late gene expression. Cells were infected in parallel with the two viruses at an MOI of 3 and collected at 48 hpi. Flow cytometry analyses showed that the fluorescence signal of both GFP and mCherry in vAcΔlef10/lef10p:lef10 1-65 -GFP/p10p:mCherry infected cells were slightly lower than the cell with vAcΔlef10/lef10p: lef10-GFP/p10p:mCherry (Fig 4A right), indicating that LEF-10 1-65 could be expressed at a lower level under the control of its intrinsic promoter or less stable than the wild-type LEF-10 and it only partially recovered the function of full-length LEF-10 (Fig 4A left).  Sf9 cells were respectively infected with vAcΔlef10/lef10p:lef10-GFP/p10p:mCherry or vAcΔlef10/lef10p:lef10 1-65 -GFP/p10p:mCherry and analyzed at 48 hpi. The expression of LEF10-GFP driven by lef-10 promoter was detected in the FL1 channel, and mCherry expression controlled by the p10 promoter was detected in the FL2 channel. The schematic diagram of expression cassette is at the bottom. (B) Fluorecence microscopic analysis of the correlation between LEF-10 aggregation and the expression of mCherry under the control of p10 promoter. When the LEF-10 aggregation was observed, the red fluorescence was almost undetectable. The cells were analyzed at 48 hpi. (C) Flow cytometry analysis of the effect of LEF-10 C-terminal truncation on LEF-10 over-expression. Sf9 cells were respectively infected with vAcΔlef10/ actinp-p10p:lef10-GFP and vAcΔlef10/actinp-p10p:lef10 1-65 -GFP, and the expression of the GFP fusion proteins, controlled by tandem chicken actin promoter and p10 promoter, was detected in the FL1 channel. The schematic diagram of expression cassette is at the right Observation of vAcΔlef10/lef10p:lef10-GFP/p10p:mCherry infected Sf9 cells by fluorescence microscopy at 48 hpi revealed that cells which formed punctate dots could hardly produce detectable red fluorescence signals (Fig 4B). In contrast, mCherry was abundantly expressed in the cells with LEF10-GFP evenly distributed (Fig 4B). This observation indicated that aggregation of LEF-10 down-regulated the expression of viral very late gene p10. Also, we observed that LEF-10 formed punctate dots predominantly located in the cytoplasm while non-accumulated LEF-10 mainly located in the nucleus (Fig 4B), indicating that LEF-10 functioned in the nucleus and aggregation of the protein in the cytoplasm probably depleted soluble LEF-10 and impeded the transportation of the protein to the nucleus.
Observed by microscopy, over-expressed LEF-10 aggregated more frequently than the LEF-10 expressed under the control of its intrinsic promoter in Sf9 cells but LEF-10 1-65 did not aggregate to form punctuate dots even when over-expressed (Fig 3B and 3C). To further investigate the relationship between LEF-10 aggregation and its activity, cells infected with vAc-Δlef10/actinp-p10p:lef10-GFP and vAcΔlef10/actinp-p10p:lef10 1-65 -GFP (the same viruses as in Fig 3B and 3C), respectively over-expressing wild-type LEF-10 and truncated LEF-10 1-65 , were further analyzed by flow cytometry. The expression level of LEF10-GFP in vAcΔlef10/ actinp-p10p:lef10-GFP infected cells was obviously lower than that in vAcΔlef10/actinp-p10p: lef10 1-65 -GFP infected cells (Fig 4C), and this result was confirmed by semi-quantitative RT-PCR (Fig 4D bottom) at the transcriptional level, but was contrary to the protein expression levels under the control of lef-10 promoter ( Fig 4A). As LEF-10 1-65 was detected at a lower level than LEF10-GFP when the protein expression was controlled by the weak lef-10 promoter, these data suggested that the aggregation of over-expressed LEF10-GFP, but not LEF-10 1-65 , contributed to the down-regulation of LEF-10 function involved in the very late gene expression. Semi-quantitative RT-PCR analyses of other five late genes, pp78/83 [29], vp39 [30], sod [31], odv-e56 [32] and vp80 [33], indicated that LEF-10 aggregation affected the transcription of these late genes as well as the very late p10 promoter controlled gene (Fig 4D).

Discussion
LEF-10, the smallest late expression factor of baculovirus, contains only 78 amino acids and was originally identified in transient-expression assays as an essential factor for AcMNPV late promoter transcription [34]. In a previous report, we found that AcMNPV LEF-10 could form very stable SDS-resistant complexes [25]. A recent study using gene knockout bacmid has shown that BmNPV LEF-10 is indispensable for the viral DNA replication and viral gene transcription [23]. However, so far, LEF-10 has not yet been extensively studied, and the structure and the specific role of the protein in virus infection still remains unknown.
Here, using the similar strategy in BmNPV [23], we knocked out lef-10 in AcMNPV bacmid and generated BacmidΔlef-10. The BacmidΔlef-10 lost the ability to produce infectious virus and re-introduction of lef-10 at the polh locus repaired its function in virus infection, demonstrating that lef-10 is also essential for AcMNPV replication. Although no typical nuclear localization signal (NLS) was found in LEF-10 sequences, nuclear localization of the protein was observed in most virus infected Sf9 cells. Its nuclear localization is consistent with its classification as a regulation factor involved in late gene expression, and also suggests the protein's probable role in late gene transcription.
site. (D) Analysis of the expression of five late genes, pp78/83, vp39, sod, odv-e56 and vp80 by semi-quantitative RT-PCR, in parrallel with the very late p10 promoter driven lef10-GFP. Insect cell β-actin was used as internal reference. The left sample was from vAcΔlef10/ actinp-p10p:lef10-GFP infected cells, and the right sample was derived from vAcΔlef10/actinp-p10p:lef10 1-65 -GFP infected cells. doi:10.1371/journal.pone.0154835.g004 Using truncated LEF-10 to repair BacmidΔlef-10, we defined the shortest functional LEF-10 moiety in the amino acids 1-48. Sequence alignment of AcMNPV LEF-10 with 28 NPVs identified three relatively conserved regions (C1, C2 and C3) in the protein (Fig 5). Our data showed that removal of part of the C1 region containing two conserved amino acids (L12 and I13) in LEF-10 14-78 caused the disruption of its function in virus infection. Similarly, when the C2 region was partially removed (LEF-10 1-47 ), repaired virus could not be generated. In contrast, the C3 region can be completely removed without affecting the virus survival. All these data indicate that the conserved C1 and C2 regions are important for the function of LEF-10 while the C3 region of LEF-10 is dispensable for AcMNPV replication. Further studies are needed to investigate whether it is the same case for other baculoviruses.
A number of proteins, including amyloid β-protein (Aβ) [35,36], sup35 [37], TAR DNAbinding protein (TDP-43) [38], α-synuclein [39] and mitochondrial protein MAVS [40], have been reported to be able to form SDS-resistant aggregates. However, their aggregated forms can be disrupted and seem not so stable as LEF-10. Although different expression and purification strategies have been tried, we are still unable to obtain purified LEF-10 monomers. Immunization of rabbit with the purified LEF-10 complex also failed to induce significant immune responses against LEF-10. These intriguing properties bring difficulties to further characterize this protein.
We believe that the aggregation nature of LEF-10 plays some role in virus life cycle. Using over-expressed LEF-10 and LEF-10 1-65 , we found that LEF-10 aggregation down-regulated the very late gene expression controlled by p10 promoter. When LEF-10 was sequestered in punctate dots which surrounded the nucleus, the expression level of mCherry driven by the p10 promoter was almost undetectable (Fig 4B). It seems that LEF-10 aggregation impeded its transportation from the cytoplasm to the nucleus, and therefore blocked its function in the activation of late gene expression. Based on these observations, a model for the relationship between LEF-10 aggregation and its function was established and illustrated in Fig 6. Protein aggregation is a biological phenomenon mainly caused by protein folding or misfolding. Protein aggregation is often harmful and has been implicated in a wide variety of diseases. In Saccharomyces cerevisiae, aggregation of Sup35p will lead to the loss of its function as the release factor [41]. In a Drosophila model, cytoplasmic aggregation of huntingtin protein blocks axonal transport [42]. In the present study, we showed that LEF-10 aggregation downregulated the expression of some late/very late genes. Sequence analysis showed that 47% amino acid residues in the LEF-10 sequence are hydrophobic, and most of the conserved residues, especially those in the C1 region, are hydrophobic amino acids (Fig 5). When the protein misfolding occurs, these hydrophobic amino acids may be largely displayed on the protein surface, causing the aggregation of LEF-10. It is an interesting question for further investigation whether these conserved hydrophobic amino acids are positively selected during virus/host coevolution.