Identification and Characterization of the Spodoptera Su(var) 3-9 Histone H3K9 trimethyltransferase and Its Effect in AcMNPV Infection

Histone H3-lysine9 (H3K9) trimethyltransferase gene Su(var) 3-9 was cloned and identified in three Spodoptera insects, Spodoptera frugiperda ( S . frugiperda ), S . exigua and S . litura . Sequence analysis showed that Spodoptera Su(var) 3-9 is highly conserved evolutionarily. Su(var) 3-9 protein was found to be localized in the nucleus in Sf9 cells, and interact with histone H3, and the heterochromatin protein 1a (HP1a) and HP1b. A dose-dependent enzymatic activity was found at both 27 °C and 37 °C in vitro, with higher activity at 27 °C. Addition of specific inhibitor chaetocin resulted in decreased histone methylation level and host chromatin relaxation. In contrast, overexpression of Su(var) 3-9 caused increased histone methylation level and cellular genome compaction. In AcMNV-infected Sf9 cells, the transcription of Su(var) 3-9 increased at late time of infection, although the mRNA levels of most cellular genes decreased. Pre-treatment of Sf9 cells with chaetocin speeded up viral DNA replication, and increased the transcription level of a variety of virus genes, whereas in Sf9 cells pre-transformed with Su(var) 3-9 expression vector, viral DNA replication slow down slightly. These findings suggest that Su(var) 3-9 might participate in the viral genes expression an genome replication repression during AcMNPV infection. It provided a new insight for the understanding virus–host interaction mechanism.


Introduction
Post-translational modifications on N-terminus of core histones, such as methylation, acetylation, phosphorylation, ubiquitination and glycosylation, etc., could affect the affinity between histones, DNA and a variety of protein factors, and change the status of chromatin compaction and gene expression [1,2]. Site-and state-specific methylations of histone Lysine or Arginine are catalyzed by a family of proteins containing SET domain, which was first identified in Drosophila Su(var) 3-9, Enhancer of zeste and Trithorax [3], from which the acronym is derived. The SET domain is highly conserved among most histone methyltransferases (HMTs). The Cterminal half of SET contains H(x 2 ) NHSC and GE(x 5 ) Y motifs, which was defined as the catalytic core [4,5]. It is often flanked by one or both of the functionally associated domains named preSET and postSET [3,4]. The preSET, which is also known as SAC (SET-domain-associated cysteine-rich region), interacts with surfaces of core SET to stabilize the structure [3,4], while the postSET contributes to the binding of methyl donor S-adenosylmethionine (SAM) and the methylation [6]. Different histone Lysine methyltransferases (HLMTs) have variable activities to catalyze Lysine methylations at different site (H1K26, H3K4, H3K9, H3K27, H3K36, H3K79 and H4K20), and to different methylation statuses (mono-, di-and trimethylation).
Virus infection usually affects multiple biological processes of the host, which may activate host antiviral responses in return. In the perspective of cellular epigenetic regulation, it is known that the genomes of many DNA viruses, including herpesviruses, adenoviruses, hepatitis B virus, are subject to histone binding, and form chromatin-like structure in the host nucleus. On the other hand, virus may make the use of epigenetic mechanisms to ensure their replication [19][20][21][22][23]. The insect baculovirus Autographa californica multiple nucleopolyhedrovirus (AcMNPV) is known to induces a series of changes in the host cells Sf9, such as cellular cytoskeleton alternation, cell cycle arrest at G2/M, cellular gene transcription inhibition and protein synthesis shut-off, etc. [24,25]. It has been suggested that AcMNPV genome also has the nucleosome structure in infected cells [26], and chemical inhibitors of HDAC and DNMT could affect virus genome replication and gene transcription [27][28][29]. However, the details of AcMNPV-host interaction at the epigenetic level are largely unknown.
In the current study, we identified Su(var) 3-9 from three Spodoptera insects, S. frugiperda, S. exigua and S. litura. The splice variant of Su(var) [3][4][5][6][7][8][9]eIF2γ, and two of the three HP1 variants, HP1a and HP1b was also identified in S. frugiperda. The expression and subcellular localization of Su(var) 3-9, and its interaction with Histone H3 substrate, HP1a and HP1b were investigated in Sf9 cells. The enzymatic activity of Su(var) 3-9 was evaluated in vitro and in vivo. Moreover, the effect of Su(var) 3-9 on virus genes expression and viral genome replication during AcMNPV infection was also studied. Our findings provide a new perspective for the potential roles of a HMT in AcMNPV-insect interaction.

Materials and Methods
Cell and larval culture, genome and RNA extraction, and complementary DNA generation S. frugiperda cell line Sf9 was cultured at 27 °C in TMN-FH medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 µg/ml streptomycin. S. exigua and S. litura larva were reared on an artificial diet (Baiyun Industrial Co., Ltd., Henan, China) at 25 °C.
The genomic DNA was extracted using Cellular Genome Extraction Kit (Songon, China) according to the manufacturer's instructions. Total cellular RNA was isolated using Trizol (Invitrogen, USA) for common applications, or Qiagen RNeasy Mini Kit (Qiagen, USA) for cDNA cloning, followed by DNase I digestion. For larval RNA extraction, fifth instar larvae were frozen in liquid nitrogen followed by repeats of freeze-thawing, and the frozen larvae were grinded. Complementary DNA (cDNA) was generated via reverse transcription with PrimeScript Reverse Transcriptase (Takara Bio, Dalian, China) for cDNA mapping, or PrimeScript RT Master Mix (Takara Bio) for quantification.

Sequence determination and bioinformatic analysis
The sequence of B. mori Su(var) 3-9, HP1a and HP1b (Table  S1) were used to search homologues in S. frugiperda against SPODOBASE at the website http://bioweb.ensam.inra.fr/ spodobase [30]. Expressed sequence tags (ESTs) that share maximum similarity with that of B. mori homologues, were obtained (information listed in Table S1). 5'-and 3'-rapid amplification of cDNA ends (RACE) were performed to clone the full-length cDNA from Sf9 cells using SMARTer RACE cDNA Amplification Kit (Clontech, USA) according to the manufacturer's instruction, with primers UPM and NUPM from kit, and additional primers listed in Table S2. The coding sequences of Su(var) 3-9/eIF2γ of S. exigua and S. litura were amplified with polymerase chain reaction (PCR) using common upstream primer s/eF and downstream primers sR and eR, respectively. Two fragments generated by PCR with primer pairs GeneF1/GeneR1 and GeneF2/GeneR2 were assembled to get the full length nucleotide sequence of Su(var) 3-9/eIF2γ locus within the genome context in Sf9 cells. All primer sequences above are listed in Table S3.
S. frugiperda Su(var) 3-9 was aligned with selected homologues from Genbank using Clustal W software [31]. Phylogenetic tree was generated by the Neighbor-Joining method using the alignment on MEGA5 software [32] with 1 000 bootstrap replicates. Percent amino acid identities between selected HMTs were calculated on BioEdit software [33] based on the alignment.
For immunofluorescence assay, Sf9 cells were seeded onto coverslips in 6-well culture plate at a confluency of approximate 40-50% and anchored overnight before being fixed with cold acetone for 10 min and subsequently permeabilized with 0.2% Triton X-100(v/v). Cells were incubated in blocking solution (10% goat serum and 1% BSA in PBS) for 1h, before being incubated with the primary antibody against Su(var) 3-9 (1:100) for 2 h. The cells were then incubated with tetraethyl rhodamine isothiocyanate (TRITC)-conjugated goat anti-mouse IgG (1:200) diluted with blocking solution for another 2 h. Specimens were covered onto slides with a drop of mounting medium (carbonate-buffered saline (pH8.5) with 90% glycerol (v/v)). Triplicate PBS-washing for 5 min were included in the intervals of each step. The Laser confocal microscope (model TCS-SP equipped with an Ar-Kr laser; Leica, Heidelberg, Germany) was used to examine the specimen.

Plasmid transfection, HMT inhibitor treatment and AcMNPV infection of Sf9 cells
Sf9 cells were seeded on 24-well culture plate at approximate 5×10 4 cells per well and cultured overnight before use. Cellfectin II Reagent (Invitrogen) was used for transfection experiments with quantified plasmid DNA following manufacturer's instruction. For HMT inhibitor (HMTi) experiment, appropriate concentration of Su(var) 3-9-specific inhibitor Chaetocin (Enzo Life Science, USA) [34] determined by MTT assay (Wang, et al. [35]) was added to the culture medium. For AcMNPV infection, Sf9 cells were inoculated at specified multiplicity of infection (MOI) by directly adding virus stock into medium and incubated for 4h at 27°C. The virus inoculum was then removed and replaced with fresh medium.

Nuclear protein extraction
Nuclear extract was prepared using hypotonic buffer solution (20 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 ) and Cell Extraction Buffer (Cat. no. FNN0011, Sigma) supplemented with 1mM phenylmethylsulfonyl fluoride (PMSF) and Protease Inhibitor Cocktail (Sigma) at recommended concentration, following manufacturer's instruction. A DNase I digestion step was added in order to release proteins from tightly coupled DNA-protein complex in the nuclear fraction. The final extract was quantified with Bradford method, aliquoted and stored at -80°C.

GST pull-down and co-immunoprecipitation
The recombinant protein GST-Su(var) 3-9 or GST was used as the bait to capture potential binding partners in the nuclear extract of Sf9 cells. GST or GST-Su(var) 3-9 protein was immobilized on Glutathione HiCap Matrix (Qiagen, USA) resin, and the bound resin was incubated with the nuclear extract following manufacturer's instruction. Nonspecific bound fraction was rinsed out and the bound fraction was eluted from the matrix and subjected to Western blot analysis.

HMT enzymatic assay in vitro and in Sf9 cells
HMT enzymatic assay in vitro was performed as described by Sun et al. [36] with minor modifications. Briefly, a 25 µl reaction system comprising 1 µg Histone H3.2 (NEB, USA), 160 µM SAM and His-Su(var) 3-9 protein in HMT reaction buffer (50mM Tris-HCl, pH 9.0, 5 mM MgCl 2 , 4 mM DTT) was incubated at 27 °C or 37 °C for 60 min. The reaction was stopped by adding equal amount of 2×SDS-PAGE loading buffer, and the total protein was analyzed by Western blot using specific antibody for trimethylated or non-methylated H3. For in vitro HMT inhibition assay, 5 µM Chaetocin was supplemented in the reaction mixture.
HMT activity of Su(var) 3-9 in Sf9 cells was examined using nuclear proteins of variously treated or non-treated Sf9 cells with EpiQuik Histone Methyltransferase Activity/Inhibition Assay Kit (H3K9) (Epigentek, USA) following the manufacturer's instruction.

Real-time quantitative PCR
Real-time quantitative PCR (qPCR) was used to determine the changes of host or viral transcription and viral DNA replication in AcMNPV infected Sf9. It was carried out on Mx3000p (Stratagene, USA) using SYBR Premix DimerEraser (Takara Bio). Template cDNA and DNA were prepared as described above. The relative level of transcriptional change was determined using 2 -ΔΔCt method [37]. For replication assay, the copy number of viral genome was calculated from the Ct values through a standard curve derived from serial dilutions of AcMNPV DNA stock. Primer sequences of qPCR were listed in Table S4.

DNase I-sensitivity assay
DNase I-sensitivity studies [27] were performed to evaluate alternation of chromatin compaction. Briefly, Sf9 cells were collected at specific time point post treatment, washed with PBS and resuspended in 300 µl supplemented RSB buffer (20mM Tris-HCl, pH 7.4, 10mM KCl, 1.5mM MgCl 2 , 1mM CaCl 2 , 0.5% (v/v) NP-40, 100µg/ml PMSF, 20µg/ml RNase A). The sample was then split into 50µl aliquots and divided into untreated and treated group equally. The treated group was subjected to DNase I digestion at 37° for 5 min with different concentrations of enzyme, and stopped with the stop buffer (50mM Tris-HCl, pH 8.0, 0.1 M NaCl, 1% (v/v) SDS and 100 mM EDTA). Samples were then incubated with proteinase K overnight, and DNA was extracted with phenol/chloroform and precipitated with ethanol. The percentage of DNA resistant to the nuclease was determined with qPCR with primers for Sf9 β-Actin and β-Tubulin (Table S4).

Statistical analysis
Representative results from at least two independent experiments were shown. Data from triplicate or more parallel experiments were used to calculate mean value and standard deviation.

Sequence determination of Spodoptera Su(var) 3-9/ eIF2γ
In holometabolic insects, Su(var) 3-9 and its functionally unrelated gene eIF2γ are commonly seen as two splice variants [38,39]. The partial EST of S. frugiperda eIF2γ was found in SPODOBASE [30] using the sequence of Bombyx mori homologue. 5'-and 3'-rapid RACE were carried out to obtain the full length sequence of eIF2γ transcript. Based on the obtained sequence, a 6 921 bp fragment was amplified from the genome, in which the full coding sequence of Su(var) 3-9 was identified. It was then verified by PCR with cDNA as template. The sequences of Su(var) 3-9 and eIF2γ were also identified in another two Spodoptera insects, S. exigua and S. litura. They were 100% identitical to those of S. frugiperda. The genomic structure of the Su(var) 3-9/eIF2γ locus and the splicing patterns was shown in Figure 1. There are three exons in Su(var) 3-9 ORF (1 782 bp, encoding a 593 aa protein) and nine exons in eIF2γ ORF (1 665 bp, encoding a 554 aa protein), with the first two exons shared by both splice variants. Each intron contains "GU" dyad oligonucleotide at the 5' splice site and "AG" dyad oligonucleotide at the 3' splice site. In addition, the polyadenylation (polyA) signals (AAUAAA) downstream of each ORF are also found in the genomic sequence.

Sequence analysis and phylogeny of Spodoptera Su(var) 3-9
The amino acid sequence of Spodoptera Su(var) 3-9 was aligned with SUV39 family members from Scoliopteryx libatrix (Sl), Bombyx mori (Bm), Tribolium castaneum (Tc), Apis mellifera (Am), Drosophila melanogaster (Dm) and Homo sapiens (Hs). They cover insect species from the orders Lepidoptera, Coleoptera, Hymenoptera and Diptera, apart from human being. In the neighbor-joining phylogenetic tree, ( Figure  2) SUV39 family members are well separated into each corresponding monophyletic group and the human SUV39s are distant from those of insect. The distantly related enzymes of SUV39 family, human G9a and SETDB1 are further apart from others, as expected. The three lepidopteran Su(var) 3-9s are most closely related (the identity between Spodoptera and S. libatrix Su(var) 3-9s is 74.20%, and the identity between Spodoptera and B. mori Su(var) 3-9s is 75.40%). According to the alignment of these three lepidopteran Su(var) 3-9s and the Su(var) 3-9 of Drosophila ( Figure S1), the Spodoptera Su(var) 3-9 contains all conserved domains found in its insect counterparts ( Figure 3A), including the common region with eIF2γ, the N-terminus of Su(var) 3-9, the Chromo domain, the preSET domain, the SET domain and the postSET domain [38,39]. Two highly conserved motifs, H(x 2 ) NHSC and GE(x 5 ) Y, both of which locate in the C-terminal half of the SET domain and are defined as catalytic core [4,5], are also present in Spodoptera Su(var) 3-9, as well as in other Su(var) 3-9s.

Expression and localization of Su(var) 3-9 in Sf9 cells
Western blot was performed to detect Spodoptera Su(var) 3-9 in Sf9 cells using homemade antiserum generated with bacterially expressed Su(var) 3-9. A band with apparent molecular mass slightly higher than predicated (68 kDa) was seen ( Figure 3B, left), most likely because of post-translational modifications such as phosphorylation [40]. No visible bands of Su(var) 3-9 were detected in the antigen-preabsorbed antiserum (obtained by mixing the anti-Su(var) 3-9 antiserum with excessive amount of purified Su(var) 3-9) and irrelevant mouse IgG groups, which verified the specificity of homemade antibody ( Figure 3B, middle and right). Immunofluorescence analysis indicated that Su(var) 3-9 mainly localized in the nuclei in Sf9 cells ( Figure 3C), which is consistent with its activity as HMT.

Interactions of Su(var) 3-9 with histone H3 and heterochromatin proteins HP1a/b
In Drosophila, the histone H3K9me3 serves as a signal for the binding of HP1 [10][11][12]. Besides, the N-terminus of Su(var) 3-9 can also directly interact with the Chromo-shadow domain of HP1 [13,14]. GST pull-down assay was carried out to examine those interactions in Spodoptera using GST-tagged proteins and the nuclear extract of Sf9 cells. As expected, the interactions of Su(var) 3-9 with its substrate histone H3 and HP1a /b (see Figure S2 for detection and localization of HP1a/b in Sf9 cells) were seen, whereas no detectable bands were found in the negative control with the GST tag as the bait ( Figure 3D). In addition, the specific interactions of Su(var) 3-9 and HP1a/b with histone H3 in Sf9 cells were also confirmed in the co-immunoprecipitation experiment ( Figure 3E)

Enzymatic activity assay in vitro and in vivo
In vitro enzymatic activity assay was performed to determine the ability of Spodoptera Su(var) 3-9 for histone H3K9 trimethylation. Various amount of His-Su(var) 3-9 (0.25~5 µg) was added to the reaction systems containing same amounts of H3 and SAM. As shown in Figure 4, more trimethylated H3K9 were observed with the increased amount of His-Su(var) 3-9 at both 27 °C and 37 °C. Higher enzymatic activity was seen at 27°C than at 37 °C. In addition, when Su(var) 3-9 specific inhibitor chaetocin [34] was added, significant inhibition of the enzymatic activity was seen (Figure 4). In the presence of chaetocin, the level H3K9me3 of the 5µg of Su(var) 3-9 reaction group was equivalent to that of the 1µg group, at both temperatures tested.
Cell viability assay showed that Sf9 cells could tolerate up to 500 nM chaetocin in the medium ( Figure 5A). When the cells were treated with increased dosage of chaetocin, general histone methylation level was gradually inhibited, indicating the presence of Su(var) 3-9 activity in the cell ( Figure 5B). In addition, the cellular DNA became more sensitive to DNase I at β-Actin and β-Tubulin loci, which reflected the decreased chromatin compaction ( Figure 5C). In contrast, the opposite trends of results were observed when the cells were transfected with a plasmid overexpressing Su(var) 3-9 ( Figure  5D).

Effect of AcMNPV infection on Su(var) 3-9 expression
Whether Su(var) 3-9 is involved in AcMNPV infection and virus-host interaction is studied. AcMNPV-infected Sf9 cells (MOI=20 plaque-forming units per cell) were examined for Su(var) 3-9 transcription by qPCR. As shown in Figure 6, transcriptions of cellular cytoskeleton genes, β-Actin and β-Tubulin declined dramatically 8-12 h post infection (p.i.), whereas the mRNA level of Su(var) 3-9 increased to more than threefold higher than that of the control after 12 h p.i. The transcriptions level of HP1a and HP1b remain similar to those in the control. DNase I-sensitivity assay was executed to examine whether the abnormal expression trends of Su(var) 3-9, HP1a and HP1b, in contrast with the majority of host genes that decreased their expression after virus infection, caused chromatin compaction. The results showed that the cellular DNA of infected Sf9 cells became less sensitive to DNase I than that of mock infected (Figure 7) at the loci of β-Actin and β-Tubulin genes 24 h p.i.

Influence of Su(var) 3-9 on AcMNPV genome replication and genes transcription
To evaluate the potential effect of Su(var) 3-9 on AcMNPV replication, Sf9 cells were treated with the chaetocin (500 nM) or transfected with Su(var) 3-9 overexpressing plasmid 24 h before being infected by AcMNPV (5 pfu/cell). Viral DNA replication and the transcription of representative genes, varying from immediate early to late stages, were examined with qPCR. As shown in Figure 8, AcMNPV genome replication was accelerated or delayed when treated with Chaetocin or transformed with overexpressing vector of Su(var) 3-9, respectively, although viral DNA replication achieved a similar level at the very late phase of infection in both cases. Minor increase in the expression of viral immediate early genes IE0/1 ( Figure 9A) and IE2 ( Figure 9B) were seen. Variable increases in the transcription level were also observed for either early genes, including lef1, lef2, lef3, dnapol, p35 and gp64, or the late genes p78/83 (orf1629), vp39 and p10 ( Table 1). The lef2 gene was most dramatically activated with over 50-fold enhancement in transcription at 12 h p.i., whereas in the cases of lef3 and p35, there were about 5 to 6-fold changes. The    peak of transcription enhancement appeared before or at 12 h p.i. for most genes, except P78/83 and P10, which appeared at 24h p.i.

Discussion
In the current study, the histone H3K9 trimethyltransferase gene Su(var) 3-9 was identified in three Spodoptera insects, S. frugiperda, S. exigua and S. litura. As in other holometabolic insects, Su(var) 3-9 overlaps with eIF2γ within the same genome loci, and the two genes are expressed by alternative splicing [38,39]. Phylogeny analysis revealed that Spodoptera Su(var) 3-9 is highly related to selected SUV39 family members of insects from Lepidoptera, Coleoptera, Hymenoptera, and Diptera. It is also homologous to human SUV39. It sits closely with the two other Lepidoptera Su(var) 3-9s from Scoliopteryx libatrix (also belonging to the family of Noctuidae as Spodoptera) and Bombyx mori (belonging to the  family of Bombycidae), whereas the two distantly related SUV39 family members, human G9a and SETDB1, locate farthest from others.
Among numerous HMTs for histone lysine or arginine modifications, Su(var) 3-9 has the specific catalytic capability to catalyze histone H3K9me3. This activity was confirmed for Spodoptera Su(var) 3-9 in vitro and in Sf9 cells. Spodoptera Su(var) 3-9 exhibits a dose-dependent enzymatic activity at both 27 °C and 37 °C, but higher activity was seen at 27 °C in the lower enzyme concentration groups in vitro, coinciding with the fact that 27 °C is close to the optimal temperature for the insects.
Spodoptera Su(var) 3-9 was also proved to interact with histone H3 and HP1a/b, and the presence of Su(var) 3-9 specific inhibitor Chaetocin or Su(var) 3-9 overexpressing vector change the chromatin compaction in Sf9 cells accordingly. This was consistent with previous studies in other species that histone H3K9me3 catalyzed by Su(var) 3-9 may serve as a signal for HP1 binding and HP1 can subsequently recruit more Su(var) 3-9 to modify neighboring histones, thus causing the formation and diffusion of heterochromatin [13].
It has been reported that AcMNPV infection may provoke multiple host responses including general down regulation of cellular mRNA level and shut-down of host protein synthesis [24,25]. This was supported by the results of DNase I sensitivity assay, which showed that host chromatin became more compacted after AcMNPV infection. Surprisingly, the transcription of Su(var) 3-9 increased in AcMNPV-infected Sf9 The genomes of a variety of DNA viruses are subjected to nucleosome assembly and are associated with histone modifications, which may inevitably take effect on viral life cycle. Even the acutely replicating DNA viruses such as Adenoviruses, which encode their own DNA polymerase and viral DNA assembly proteins, do not escape chromatin assembly. This might be helpful for viruses for their proper temporal control of viral gene transcription and genome stability during nuclear entry and the early stages of virus replication. Taken together with the previous reports that AcMNPV genome also has nucleosome-like structure during infection, and that several epigenetic drugs such as TSA, NaBu and DAC have significant effect on viral genome replication and gene transcription [26][27][28][29], it is reasonable to believe that changes in histone modification may have a role in AcMNPV genome replication and gene transcription. In this study we demonstrated that Su(var) 3-9 specific inhibitor Chaetocin or transient expression of Su(var) 3-9 could activate or repress AcMNPV genome replication and gene transcription, respectively, suggesting that Su(var) 3-9 is possibly involved in AcMNPV infection. It deserves more detailed study due to the complex epigenetic regulation network. Figure S1. Alignment of amino acid sequence of Su(var) 3-9s from S. frugiperda (Sf), Scoliopteryx libatrix (Sl), Bombyx mori (Bm), and Drosophila melanogaster (Dm). Identical residues are dark black shadowed with white font. Similar residues are gray shadowed. The regional subdivisions of common region with eIF2γ, N-terminus of Su(var) 3-9 Chromo domain, preSET, SET and postSET are indicated within black framed below the sequence. The catalytic core motifs of H(x 2 ) NHSC and GE(x 5 ) Y are marked with red asterisks. (TIF)