3H-31, A Non-structural Protein of Heliothis virescens ascovirus 3h, Inhibits the Host Larval Cathepsin and Chitinase Activities

Insect viruses are members of entomopathogens that can invade and reproduce in their selective host species and infect other host individuals. Baculoviruses are insect viruses containing circular double-stranded DNA genomes that are enveloped in rod-shaped capsids and embedded in polyhedral occlusion bodies (Federici 1997). Infection of permissive insect hosts by baculoviruses often results in liquefaction, facilitating the dispersal and transmission of progenies that could then infect other hosts (Blissard and Theilmann 2018). Owing to their restricted infectious host range, baculoviruses are safe for non-target organisms and the environment; hence, baculoviruses are consequently developed into biopesticides (Huber 1986; Erlandson 2008; Martignoni 1999; Lord 2005; Lacey et al. 2008; Lacey and Shapiro-Ilan 2008). Besides, baculoviruses are widely employed as eukaryotic expression systems and gene therapy vectors (Keeler and Flotte 2019; Gouma et al. 2020).

Ascoviruses are insect viruses that also contain doublestranded DNA genomes, but their allantoic-shaped virions are embedded in flexible and porous vesicles, and ascoviruses do not have compact structural occlusion bodies (Asgari et al. 2017; Cheng et al. 1999). Furthermore, ascoviruses are humorally transmitted viruses that rely on the oviposition of parasitic wasps to commit their transmission between other host individuals (Federici et al. 1990; Hamm et al. 1986; Li et al. 2016). Instead of climbing up to the top of the host plants and liquefaction, as exhibited in the baculovirus-infected caterpillar (Bhattarai et al. 2018; Hoover et al. 2011), the ascovirusinfected larva had one typical symptom—the color change in their hemolymph from lucent clear into turbid milky white (Govindarajan and Federici 1990). No melting was detected in the larval cadavers that died from the infection of ascoviruses, which might be the reason why ascoviruses were not found until 1983 (Federici 1983), centuries later than the discovery of baculoviruses.

The infection characteristics of ascoviruses have been reported to be apoptosis-like processes, in which the ascoviruses manipulates host cellular apoptosis by coding a caspase-like protein, and the ascovirus utilizes the host apoptotic bodies to generate their vesicles (Federici et al. 2009). Thus, apoptotic-associated genes encoded by the ascoviruses have been enthusiastically characterized in the early ten years since its discovery (Bideshi et al. 2005; Asgari 2007; Hussain and Asgari 2008; Smede et al. 2009). Moreover, the structural protein of the ascovirus were also major topics of debate from the proteomics determination to functional analysis of specific structural proteins, several structural proteins of ascoviruses were characterized (Tan et al. 2009; Chen et al. 2018; Stasiak et al. 2003; Salem et al. 2008; Zhao et al. 2019a). However, compared with approximately two hundreds of open reading frames (ORFs) that are commonly encoded by ascoviruses, the understanding of these proteins is far from sufficient. One of the reasons for the inadequate studies on the genes of ascovirus might be the lack of artificial chromosomes of the ascovirus, which have been widely used in the deletion of specific genes in baculoviruses.

Heliothis virescens ascovirus 3h (HvAV-3h) is a member of the ascoviruses that infect several noctuid (Lepidoptera) species, including Spodoptera exigua, S. litura, S. frugiperda, Helicoverpa armigera, Mythimna separata, among others (Huang et al. 2012; Li et al. 2013; Yu et al. 2020a). The host larva infected with HvAV-3h will stay at the larval stage until death, and the chitinase activity is significantly influenced and therefore cannot ensure normal growth and development (He et al. 2020). Unlike other insect viruses, HvAV-3h can use melanization of the host larvae to help virus replication (Li et al. 2021). One hundred and eighty-five ORFs were predicted to be encoded by the genomic DNA of HvAV-3h (Huang et al. 2017). The 3H-117 and 3H-21 were revealed to be two structural proteins of the virions of HvAV-3h, which have similar functions as the major capsid protein (MCP) (Zhao et al. 2019a, 2019b). No other functional genes of HvAV-3h have been characterized.

To understand the function of 3h-31 (orf31 of HvAV-3h), the bacmid of Autographa californica nucleopolyhedrovirus (AcMNPV) was employed as an overexpression system to express the 3h-31 gene by repairing enhanced green fluorescent protein gene (egfp)-fused 3h-31 under p10 promoter into the AcMNPV genome. Interestingly, 3h-31 was associated with the regulation of chitinase and cathepsin activity in host larvae. The function was further confirmed by the RNA interference (RNAi) of 3h-31 in the HvAV-3h infected S. exigua and H. armigera.

Laboratory colonies of beet armyworm (S. exigua) and cotton bollworm (H. armigera) were maintained with artificial diets at 27 ℃ ± 1 ℃ and a 16-h light/8-h dark photoperiod (Yu et al., 2020b). The SeFB cells (IOZCASSpexII-A) (Zhang et al., 2009) derived from the fat body of S. exigua was kindly provided by Dr. Qilian Qin, and maintained in TMN-FH insect cell culture (Sigma, USA) supplemented with 10% fetal bovine serum (FBS) (GIBCO, AUS) at 27 ℃. The Ha-E cells derived from the embryo of H. armigera were generously provided by Prof. Jianhong Li (College of Plant Science & Technology, Huazhong Agricultural University) and maintained in TMN-FH insect cell culture supplemented with 10% FBS (GIBCO) at 27 ℃.

Heliothis virescens ascovirus 3h was isolated in China and stored at -20 ℃ in our laboratory (Huang et al. 2012). For the ascovirus inoculation, the HvAV-3h-containing hemolymph was used according to the method described by Yu et al. (2018b). Escherichia coli strain DH10B, which contained the AcMNPV genome as a large plasmid (bMON14272), was generously provided by Dr. Zhihong Hu (Wuhan Institute of Virology, Chinese Academy of Sciences) and was used to construct the 3h-31 recombinant AcMNPV.

To infect the Ha-E or SeFB cells, the ascovirus-containing cell culture medium was prepared first (Yu et al. 2019b). The HvAV-3 h-containing hemolymph collected from the S. exigua larvae at 7 days post-infection (dpi), with a virion content of about 1.1 × 10¹¹ virion copies per mL, was added into FBS-free cell culture medium with a dilution of 1,000. After sterilization with filtration (0.22 μm), FBS was added to the ascovirus-containing cell culture medium (1.1 × 10⁸ virion copies per mL) to a final concentration of 10%. Hemolymph collected from healthy S. exigua larvae were diluted and sterilized in the same way and used as a negative control (mock-infected control) in the following assays.

The Ha-E or SeFB cells were seeded in 6-well plates to generate 10⁵ cells per well. One milliliter of the prepared HvAV-3h-containing medium or the negative control medium was added to each well. The supernatant of each well was replaced with fresh culture medium after 1 h of incubation, and this point was set to 0 dpi. The Ha-E or SeFB cells infected with HvAV-3h were collected at 0, 3, 6, 12, 24, 48, 72, and 96 h. The total RNA and protein content were extracted from the collected cell samples using TRI reagent (MRC, USA) according to the manufacturer's instructions. The cDNA was synthesized using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) with 1.0 μg of the total RNA extracted from larvae or cells as the template. The primer pair 3h-31- F(3'-GGATCCATGAACGGACACGTCGAG-5', BamHI site underlined)/3h-31-R (3'-CTCGAGTTATTCGATTACACAAGT-5', XhoI site underlined) were used to detect the temporal transcription of 3h-31 by PCR with the synthesized cDNA samples. The primer pair SeGapdhF(5'-ATGTCTAAAATCGGTA-3') and SeGapdh-R(5'- ATCCTTGGTCTGGATG-3') was used to detect the transcription of the glyceraldehyde-3-phosphate dehydrogenase gene (gapdh) of S. exigua, which was used as a reference gene in the treated S. exigua larvae or SeFB cells. The primer pair HaGapdh-F(5'-ATGTCCAAAATCGGTATCAA-3') and HaGapdh-R(5'-TTAATCCTTGGTC TGGATGTA-3') was used to detect the transcription of gapdh from H. armigera, which was used as a reference gene in the treated H. armigera larvae or Ha-E cells.

The specific antiserum of the 3H-31 protein was prepared prior to the Western blotting analysis (the expression and purification of 3H-31 is illustrated in the Supplementary Figure S1 and S2). The total protein of HvAV-3hinfected cells was extracted from the organic phase of the above RNA extraction tubes, according to the manufacturer's instructions for the TRI reagent (MRC, USA). The resulting protein samples extracted from the ascovirusinfected larvae or cells were mixed with 5 × loading buffer and incubated in boiled water for 10 min. After separation using a 12% SDS-PAGE system, the proteins were transferred to a nitrocellulose membrane. The prepared 3H-31 polyclonal antiserum (1:3000) was used as the primary antibody to detect the expression of 3H-31. A polyclonal antibody against ascovirus MCP (1:3000) was used to detect the expression of MCP to confirm the infection of ascoviruses. The noctuid insect-specific GAPDH antibody (1:4000) was used as a reference antibody (Yu et al. 2018a). Horseradish peroxidase (HRP)- conjugated goat anti-rabbit (1:5000) (Millipore, USA) were used as the secondary antibodies. The proteins were visualized using ClarityTM Western ECL Substrate (BioRad, USA).

Fresh HvAV-3h-containing hemolymph was collected from the S. exigua larvae at 7 dpi and suspended in 10 volumes of ice-cold TE buffer (10 mmol/L Tris, 1.0 mmol/L EDTA, pH 7.4) with complete, EDTA-free, Protease Inhibitor Cocktail (Roche, SUI). After ultrasonication, the resulting mixture was loaded onto a prepared 15%-35%- 55% (w/v) discontinuous sucrose gradient, followed by centrifugation (72,100 ×g for 1.5 h, 4 ℃). The virion bands (between the 15%-35% sucrose gradient) were collected and washed three times with 10 volumes of icecold TE buffer (110,000 ×g for 1.5 h, 4 ℃) and finally resuspended in 1/10 volumes of TE buffer and stored at -80 ℃. The purified virions were negatively stained and observed by transmission electron microscopy (H-7650, Hitachi, Japan) as described by Zhao et al.(2019a, b).

The total protein of the purified HvAV-3h virions was extracted using RIPA lysis buffer (Solarbio, China) according to the manufacturer's instructions. The hemolymph protein collected from the healthy or HvAV-3h infected S. exigua larvae were used as controls. The protein samples were separated by SDS-PAGE followed by transfer onto the nitrocellulose (NC) membrane. Polyclonal antiserum of the 3H-31, MCP, and GAPDH proteins was used to identify the expression and location of 3H-31.

In order to understand the function of 3h-31, a recombinant AcMNPV that contained 3h-31 and egfp was constructed. A Bst1107I and BclII sited a 1865-bp fragment, including polyhedrin with its native promoter, and an SV40 terminator as well as a p10 promoter and HSV tk polyA terminator with several restriction enzyme sites between them were synthesized. Two nucleotides in the polyhedrin sequence were modified to silence the native containing the HindIII and EcoRI sites. The synthesized ph-Pp10 fragment was digested with Bst1107I and BclII and ligated with the pFastBac HTb vector (Invitrogen, USA), which was also digested with Bst1107I and BclII to generate the phPp10-HTb vector. The egfp was then subcloned into the SalI/XhoI sites of the ph-Pp10-HTb vector to generate the donor plasmid (ph-Pp10-egfp-HTb), which can express egfp under the control of the p10 promoter. The 3h-31 gene was inserted into the BamHI/XhoI sites of the ph-Pp10- egfp-HTb, which was upstream of egfp. The resulting phPp10-31-egfp-HTb plasmid was transformed into DH10Bac cells to generate the polyhedron-repaired and 3h-31-fused egfp-inserted AcMNPV bacmid (AcMNPV- 31) according to the manufacturer' s instructions (Invitrogen, USA). The ph-Pp10-egfp-HTb vector was transformed into DH10Bac cells to generate a control AcMNPV bacmid (AcMNPV-wt), which can express EGFP.

One microliter of the constructed AcMNPV-wt or AcMNPV-31 bacmid DNA was transfected into one 35-mm-diameter well with cultured SeFB cells using the X-tremeGENE HP DNA Transfection Reagent (Roche). The cells and their supernatants were collected at 0, 3, 6, 12, 24, 48, 72, and 96 h post transfection (hpt), respectively. The supernatant samples that contained budded viruses (BVs) were used to analyze the viral titer using the limiting dilution method, and the 50% tissue culture infective dose (TCID₅₀) value was calculated using the method described by Lei et al. (2020). The collected cells were used to extract the total DNA with SteadyPure Universal Genomic DNA Extraction Kit (Accurate Biology, China), and the extracted DNA was used as a template for qPCR (Yu et al. 2019b). The viral DNA content in each sample was calculated, and the viral DNA replication curve was established. The difference of viral DNA contents and BVs production at each tested time of points between AcMNPV-wt and AcMNPV-31 infected Sf9 cells were compared by Student's t test. The comparisons got a P value below 0.05 during the t tests indicated a significant difference of viral DNA contents or BVs production between AcMNPV-wt and AcMNPV-31 infected Sf9 cells, and were marked with asterisk. The general linear model (GLM) in SPSS 22.0 (SPSS Inc., Chicago, IL, USA) was employed to analyze the difference between the different virus strains.

The SeFB cells transfected with AcMNPV-wt or AcMNPV-31 were collected at 72 hpt and washed three times with PBS (centrifuging at 1500 ×g for 10 min). The cells were then fixed with 2.5% glutaraldehyde overnight, followed by dehydration, embedding, section, and staining according to the description of Xu et al. (2008). Samples were examined using a Hitachi H-7650 TEM with an accelerating voltage of 80 kV.

To increase the scale of virus production, 4 μL of each BV supernatant (produced from AcMNPV-wt- or AcMNPV- 31-infected SeFB cells) with a titer of 10⁵ TCID₅₀ mL^-1 was injected into the coelom of the fourth instar S. exigua larva. Larvae injected with SFX-insect culture medium were used as controls. The injected larvae were individually cultured in 24-well insect culture boxes until they were pupated or dead. The larval cadavers that died from the virus infection were collected, and the viral occlusion bodies (OBs) were purified according to the description of Yu et al. (2015a). The purified OBs of AcMNPV-wt or AcMNPV-31 were coated on coverslips and coated with gold. The morphology of the OBs was examined using a Hitachi 5X-300000X scanning electron microscope (Hitachi, Japan).

The lethal dosage of AcMNPV-31 and AcMNPV-wt against S. exigua larvae were tested by a droplet-feeding bioassay (Hughes et al. 1986; Yu et al. 2015a). Third-instar S. exigua larvae were starved for 12 h, followed by feeding with the suspension of AcMNPV variants' OBs (5, 10, 50, 100, 500, 1000, and 5000 OBs/larva) mixed with 5% sucrose and 3% blue food color (Shaanxi TOP Pharm Chemical Co., Ltd, China). The larvae that ingested the whole droplet within 10 min were transferred into the cells of 24-well insect culture boxes and supplied with fresh artificial diet dots. Tested larvae were maintained at 27 ℃ and monitored daily. Dead larvae were recorded and removed daily until all the tested larvae have died or pupated. Virus-free sucrose and blue food color mixture with fed S. exigua larvae was used as CK. Thirty larvae were used at each dosage level, and four repeats were performed with each AcMNPV variant. The median lethal time (LT₅₀) values of AcMNPV variants were tested with third instar S. exigua larvae fed with the lethal dose for 90% of the population (LD₉₀), and three repeats were performed. The data of the AcMNPV-wt or AcMNPV-31 were analyzed using the Probit analysis in SPSS 22.0 (SPSS Inc., Chicago, IL, USA).

The artificial diet consumption and larval weight gain of the virus-infected larvae were analyzed with LD₂₀, LD₅₀, and LD₉₀ dosage-fed third instar S. exigua larvae. The virus-fed larval individuals were weighed daily, and their diet consumption was calculated by the daily weight change of the supplied diet. Healthy larvae fed with virusfree diets were used as controls. Thirty larvae were used for each dosage treatment. The daily weight gain and diet consumption of the tested larvae for various AcMNPVs were compared using the general linear model (GLM) in SPSS 22.0 (SPSS Inc., Chicago, IL, USA).

Healthy S. exigua, AcMNPV-wt-infected, and AcMNPV- 31-infected larvae were transversely cut into two to three segments to expose their cross-sections. The fourth to eighth segments were fixed with tenfold volumes of polyoxymethylene-sucrose solution (4% polyoxymethylene, 5% sucrose in phosphate buffered saline, pH 6.4) for 24 h at 4 ℃. The freshly fixed tissue parts were dehydrated with a series of graded ethanol series (30%, 50%, 70%, 80%, 90%, 95%, and 100%) for 15 min each. After an additional absolute ethanol dehydration, the tissues were treated with xylene/ethanol 1/1 (V/V) for 30 min, xylene/ethanol 2/1 (V/V) for 30 min, and absolute xylene for 60 min to permeabilize the samples. For infiltration, the samples were immersed in a tenfold volume of paraffin/xylene 1/1 (V/V) for 30 min, paraffin/xylene 2/1 (V/V) for 30 min, and absolute paraffin for 30 min. The tissues were embedded externally with paraffin and used for sectioning to create 3-8 lm transverse section slices. The slices were pasted onto slides using a slice paste reagent (AngYuBio, China), followed by baking for 60 min at 70 ℃. The tissue specimens were stained with hematoxylin and eosin (H&E) according to the manufacturer's instructions in H&H Informational Primer (Sigma, USA). The slides were observed by reversed microscopy (Vert.A1, ZEISS).

AcMNPV-wt or AcMNPV-31 was fed to the third instar larvae (S. exigua) at an LD₉₀ dosage. The virus-infected larvae were collected at 1, 2, 3, 4, and 5 dpi and homogenized with PBS (approximately 20 mg of tissue per 1 mL buffer). After centrifugation (12,000 ×g) for 5 min at 4 ℃), the supernatant was transferred into another tube and used as an enzyme extract in the following cathepsin and chitinase activity assays. Six larval individuals were collected at each set time point, and healthy larvae were used as controls.

Before the enzyme activity determination, the protein content was detected with the Pierce^TM BCA Protein Assay Kit (Thermo Fisher Scientific). The cysteine proteinase activity (cathepsin L) was determined according to the method described by Riemann et al. (1982). Fifty microliters of the enzyme homogenate were mixed with 50 μL assay buffer (50 mmol/L potassium acetate, 0.1% Triton X-100, 5 mmol/L 2-mercaptoethanol, pH 5.0) or inhibitor (20 μmol/L Z-Phe-Phe-CHN₂, 2 μmol/L pepstatin) and preincubated at 1 ℃-4 ℃ for 10 min. One hundred microliters of the substrate solution (2% azocasein, 6 mol/L urea, pH 5.0) was added, followed by incubation at 37 ℃ for 60 min. After being cooled for 2 min at 4 ℃, 200 μL ice-cold trichloroacetic acid (10%, w/v) was added. After centrifugation at 100,000 ×g for 1 min, the supernatants were transferred into 96-well plates and read at 366 nm. Blanks were prepared by adding trichloroacetic acid to the enzymes or buffer before adding the substrate. The cathepsin L activity was calculated as: ((ΔE of the control—ΔE of the inhibited probe)/ (protein content)) (ΔE = extinction difference after incubation and blank subtraction).

Prior to chitinase activity determination, colloidal chitin was prepared according to the method described by Yu et al. (2015b). One hundred microliters of the enzyme homogenate was mixed with 100 μL of chitin colloid, and incubated at 37 ℃ for 1 h. After centrifugation (10,000 ×g for 5 min, 4 ℃), 100 μL of the supernatant was transferred into a new tube, and 20 μL potassium tetraborate (0.8 mol/L) was added and mixed thoroughly. The mixture was incubated in boiling water for 3 min, followed by immediate chilling in flowing water. Six hundred microliters of 4-dimethylaminobenzaldehyde (DMAB, 10 g DMAB dissolved in 100 mL glacial acetic acid containing 12.5% of 10 mol/L hydrochloric acid) was added and incubated at 37 ℃ for 20 min, and the supernatants were transferred into 96-well plates and read at 585 nm. An N-acetylglucosamine (N-GlcNAc) standard curve was used to calculate the N-GlcNAc content produced from the chitin colloid digested by the larval chitinase. The data of healthy larvae, AcMNPV-wt-, or AcMNPV-31-infected larvae were analyzed by two-way analysis of variance (ANOVA) and the differences between different treatment groups were compared with LSD methods with SPSS 22.0.

Healthy as well as baculovirus (AcMNPV-wt or AcMNPV-31) infected S. exigua larvae (4 dpi) were collected and used to extract the total protein as described above. The prepared protein samples were separated by 12% SDS-PAGE prior to transfer into the NC membrane. After blocking with 4% defatted milk powder containing TBST buffer (10 mmol/L Tris-Cl, 150 mmol/L NaCl, 2% Tween 20, pH 7.4), the polyclonal antibody that specifically recognizes the S. exigua chitinase (SeCHIT7 and SeCHIT11) and the chitin binding domain (CBD) containing the protein of S. exigua (SeCBD) prepared by He et al. were used to detect the expression levels of SeCHIT7, SeCHIT11, and SeCBD, respectively (He et al. 2020). To confirm the expression of 3H-31, commercial obtained His-tag Monoclonal antibody (Proteintech, CHN, Cat#66,005-1-lg) was used as primary antibody to detect the expression of EGFP fused 3H-31 in AcMNPV-31 infected S. exigua larvae or the expression of EGFP in AcMNPV-wt infected S. exigua larvae. Furthermore, a P10 specific rabbit polyclonal antibody was used to detect the viral infection in different variants infected larvae. The GAPDH polyclonal antibody was used as the reference antibody. The HRP-conjugated goat antirabbit (1:5000) (Millipore, USA) was used as the secondary antibody. The proteins were visualized using ClarityTM Western ECL Substrate (Bio-Rad, USA). Three repeats were performed for each viral treatment.

In order to detect the function of 3h-31 in the HvAV-3h infected larvae, RNA interference was performed. The dsRNA of 3h-31 (specific to 17-555 nt site of 3h-31) was synthesized using the T7 Ribomax Express RNAi System (Promega). One microliter (1 μg/μL) of the dsRNA was injected into HvAV-3h-infected H. armigera larvae (72 hpi) or HvAV-3h-infected S. exigua larvae (72 hpi), respectively. The egfp dsRNA injected ascovirus infected larvae were used as control. After injection of dsRNA for 24 h, larvae were collected and used to confirm the RNAi efficiency and to determine the cathepsin and chitinase activity changes. To confirm the RNAi efficiency, the total protein of three HvAV-3h-infected larvae and dsRNA-injected larvae were extracted using the RIPA reagent (Solarbio, China) according to the manufacturer' s instructions. The expression level of 3H-31 in different samples was detected by Western blotting with antiserum against 3H-31 prepared above. The HvAV-3h-infected larvae (1-7 dpi) as well as dsRNA-injected larvae were used to analyze the cathepsin and chitinase activity as described above. Six repeats were performed. The enzyme data were calculated and analyzed as described above using ANOVA and the differences between different treatments were compared with LSD methods with SPSS 22.0.

The predicted tertiary structure of 3H-31 is shown in Fig. 1A. With the prediction of I-TASSER, 3H-31 is mainly constructed with seven helixes, eight sheets, and ten coils. The 23, 27, 28, 94, 95, 96, 148, 149, and 293 amino acid residues were the predicted chitin binding sites (Fig. 1A). The time course of 3h-31 transcription and 3H-31 expression in HvAV-3h infected Ha-E cells or SeFB cells was analyzed by reverse transcription PCR and Western blotting using polyclonal rabbit antibodies against 3H-31 (see Supplementary Figure S1 and S2). In HvAV-3h infected Ha-E cells, the 981-bp 3h-31 specific fragment was detected from 6 hpi and persisted up to 96 hpi (Fig. 1B).

Figure 1.  Temporal expression and localization of 3H-31. A Structural prediction and possible ligand binding prediction of 3H-31. The tertiary structure prediction of 3H-31 was predicted by I-TASSER. C-score ranges from - 5 to 2, where a C-score of a higher value signifies a model with a higher confidence. Chitin (CHI, the green dot with blur skeleton) was predicted as the ligand of 3H-31, and the 23, 27, 28, 94, 95, 96, 148, 149, 293 amino acid sites were predicted to be the ligand binding sites of CaN-SubB. C-Score is the confidence score of the prediction ranging from 0 to 1, where a higher score indicates a more reliable prediction. B Temporal transcription and expression of 3h-31 in HvAV-3h-infected Ha-E cells and SeFB cells. At different time points post infection, the transcription of 3h-31, mcp, and gapdh were detected via RT-PCR with the total RNA extracted from HvAV- 3h-infected Ha-E cells or SeFB cells. The total proteins of HvAV-3hinfected Ha-E cells or SeFB cells were first separated with 12% SDS-PAGE, followed by immunoblotting assays with the prepared 3H-31, MCP, and GAPDH specific antiserums. M, marker. C Purification of HvAV-3h virions. The interface of 15%-35% sucrose was collected, washed, and coated on copper mesh, followed by negative staining and TEM observation. D Localization of 3H-31 in purified virions. The total protein of purified virions (V) and the hemolymph of HvAV-3h infected S. exigua larvae (7 dpi) (I) were resolved on 12% SDS-PAGE gels. The blots were probed with the polyclonal antiserum of 3H-31, MCP, and GAPDH to detect the localization of 3H-31, MCP, and GAPDH, respectively. M, prestained protein maker.

The mcp started to get transcribed at × hpi and remained stably transcribed at 96 hpi (Fig. 1B). A major immunoreactive band of approximately 37.3 kDa was detected at × hpi with the prepared 3H-31 antiserum, and persisted up to 96 hpi. In contrast, the immunoreactive bands of the MCP antiserum were detectable at 12-96 hpi. A slight difference in the temporal expression of 3h-31 was detected in the HvAV-3h-infected SeFB cells. The 3h-31 started to be transcribed at 3 hpi, and the immunoreactive band of 3H-31 was detected from 3 to 96 hpi in the infected SeFB cells, which was 3 h ahead of that in the HvAV-3hinfected Ha-E cells. Furthermore, the transcription of mcp was started at 6 hpi in the HvAV-3h-infected SeFB cells, which was also three hours ahead of the mcp transcription in the HvAV-3h-infected Ha-E cells. These results suggest that 3h-31 is an early gene of HvAV-3h, and SeFB was more sensitive to the infectivity of HvAV-3h compared to Ha-E cells.

In order to detect whether 3H-31 is a structural protein of HvAV-3h, the virion of HvAV-3h was purified. After centrifugation, the interfaces collected and observed by TEM between the sonically disrupted hemolymph were 15% sucrose, 15%-35% sucrose, and 35%-55% sucrose (see Supplementary Figure S3). Pure virions were obtained from the 15%-35% sucrose interface (Fig. 1C). Immunoblotting was performed with the purified virion protein samples (V) and the lysate of HvAV-3h-infected hemolymph (I), and the results are shown in Fig. 1D. The immunoreactive bands of 3H-31 antiserum, GAPDH antiserum, and MCP antiserum were detected in the HvAV-3hinfected hemolymph lysate loaded lane, but only an MCP immunoreactive band was detected in the virion proteinloaded lane. These results indicate that 3H-31 is not a structural protein of HvAV-3h.

To determine the function of 3h-31, a Bac-to-Bac insect expression system was employed to construct a 3h-31 inserted and per os infectious AcMNPV (Fig. 2A). A 1865-bp fragment (provided as Supplementary Figure S4), which contained the polyhedrin gene under its native promoter and an opposite directed p10 promoter, was synthesized (Fig. 2B-a) and subcloned it into the pFastBac HTb vector with BclII/Bst1107I digestion to form the bidirectional expression plasmid ph-Pp10-HTb (Fig. 2B-b). The egfp gene was inserted into the ph-Pp10-HTb to form the donor vector ph-Pp10-egfp-HTb (Fig. 2B-c), which was electrotransformed into DH10Bac cells to generate an EGFP-expressing control virus (AcMNPV-wt, Fig. 2B-f). Then, 3h-31 was inserted into the ph-Pp10-egfp-HTb (upstream of egfp and under the control of the p10 promoter, Fig. 2B-d), followed by transposition to generate the final recombinant viruses AcMNPV-31 (Fig. 2B-e).

Figure 2.  Construction of 3h-31 expressing and wild-type AcMNPV bacmids. A Strategy for the construction of egfp-inserted and polyhedron-repaired AcMNPV (AcMNPV-wt) and 3h-31-inserted oral infectious AcMNPV (AcMNPV-31). The polyhedrin and egfp were oppositely promoted with the polyhedrin promoter and p10 promoter, respectively. 3h-31 was inserted into the upstream of egfp with a mutation at its TAC stop code so as to fuse EGFP at the N-terminal of 3H-31. B Restriction enzyme digestion and PCR identification of the constructed vectors. (a) The synthesized ph-Pp10 fragment digested from the pUC57 vector with Bst1107I and BclII; (b) Constructed ph-Pp10-HTb vector identification by Bst1107I and BclII digestion; (c) Constructed donor plasmid digested with SalI and XhoI to detect the egfp fragment; (d) Constructed donor plasmid digested with PstI and KpnI to detect the polyhedrin fragment; (e) The PCR products from bMON14272 and AcMNPV-31 transposited clone with pUC-M13 primers. (f) The PCR products from AcMNPV-wt transposited clone with pUC-M13 primers. C Microscopic observation of transfected or infected SeFB cells. The green fluorescence expressed cells were detected in AcMNPV-wt and AcMNPV-31- transfected or -infected SeFB cells. Typical occlusion bodies were observed in AcMNPV-wt- or AcMNPV-31-infected SeFB cells. D Viral DNA replication curves of AcMNPV-wt or AcMNPV-31 established from viruses transfected with SeFB cells. E. Viral growth curves established from AcMNPV-wt- or AcMNPV-31-transfected SeFB cells. Error bars represent standard errors. The asterisks (*) represent significant difference (P < 0.05) between AcMNPV-wt and AcMNPV-31 according to Student's t test; "ns" represent no significant difference (P ≥ 0.05) between AcMNPV-wt and AcMNPV-31 according to Student's t test.

The DNA of the constructed recombinant AcMNPV strains was extracted and transfected into SeFB cells (Fig. 2C). At 96 hpt, abundant green fluorescence was detected in the AcMNPV-wt-transfected cells, while relatively weak green fluorescence was found in the AcMNPV- 31-transfected cells; similar reduced green fluorescence was detected in the AcMNPV-31-infected SeFB cells as compared to the AcMNPV-wt-infected SeFB cells. Typical occlusion bodies (polyhedra) were found in the AcMNPVwt- or AcMNPV-31-infected cells (bright observation figures). Further viral one-step growth curves and viral DNA replication curves was analyzed. Based on general line model analysis, AcMNPV-31-transfected cells had significantly decreased BV production (F = 412.55; d.f. = 1, 2; P < 0.0001, Fig. 2D) and DNA replication (F = 353.88; d.f. = 1, 2; P < 0.0001) (Fig. 2E). These results suggest that the expression of 3H-31 in SeFB cells reduces the In vivo viral infection cycle of AcMNPV.

The results described above indicate that BV production in the AcMNPV-31-infected SeFB cells was lower than that in the AcMNPV-wt-infected SeFB cells. To further analyze whether 3H-31 affects the structure of the nucleocapsid, transmission electron microscopic observation was performed with the virus-infected SeFB cells collected at 72 hpi (Fig. 3). The virogenic stroma was easily found in the nucleus of both the AcMNPV-wt- and AcMNPV-31-infected SeFB cells (Fig. 3A, 3D). Electron-dense rodshaped nucleocapsids were generated and packaged at the edges of the virogenic stroma (Fig. 3B, 3E). In addition to the normal packaged nucleocapsids (indicated by red arrows in Fig. 3), masses of lucent tubular structures (indicated by yellow arrows in Fig. 3) were observed at the surroundings of the normally generated nucleocapsids in the AcMNPV-31-infected SeFB cells (Fig. 3E). Clusters of elongated tubular structures that appear to not contain viral DNA genomes were also observed around the assembled polyhedra of AcMNPV-31 (Fig. 3F). Interestingly, the polyhedra of AcMNPV-31 were normally assembled and embedded with arrayed ODVs, and indistinguishable morphological changes were found between the polyhedra of AcMNPV-31 and that of AcMNPV-wt (Fig. 3C and 3F). These observations indicate that the insertion of 3H-31 led to incomplete packing of AcMNPV but did not affect the assembly of the occlusion body.

Figure 3.  Electron microscopic observation of AcMNPV-wt- or AcMNPV-31-transfected SeFB cells. A Portion of a cell infected with AcMNPV-wt. At 96 hpi, virogenic stroma (VS) was formed in the nucleus of the infected SeFB cells. B Normal nucleocapsids were observed at the edges of VS. C Image of normal virions with embedded polyhedra within the nucleus of AcMNPV-wt-infected SeFB cells. D Portion of a cell infected with AcMNPV-31. Virogenic stroma was found in the nucleus of the infected SeFB cells at 96 hpi. E Abundant nucleocapsids were generated at the electron-dense edges of VS in AcMNPV-31-infected SeFB cells. F Image of the assembling polyhedra produced in the nucleus of AcMNPV-31- infected SeFB cells. VS, virogenic stroma; nm, nuclear membrane (pointed by blue arrows); Nc, nucleocapsid (pointed by red arrows); OB, occlusion body (pointed by green arrows).

Further morphological comparison of the polyhedra was performed by scanning electron microscopic observation with the occlusion bodies purified from the larval cadavers infected with AcMNPV-wt or AcMNPV-31 (Fig. 4A). The occlusion bodies of AcMNPV-wt were typical polyhedral with smooth surfaces. The occlusion bodies of AcMNPV- 31 ranged in size and shape, and obvious grooves with empty ODVs were found on the surface of the polyhedra. These results indicate that the maturity of the AcMNPV-31 occlusion bodies is disrupted in the virus-infected larvae.

Figure 4.  Bioassays of AcMNPV-wt and AcMNPV-31 against third instar S. exigua larvae. A. Scanning electron microscopic observation of the purified OBs of AcMNPV-wt or AcMNPV-31. B. Survival curve of third instar S. exigua larvae infected with AcMNPV-wt or AcMNPV-31. The tested larval individuals were fed with 5, 10, 50, 100, 500, 1000, 5000 occlusion bodies (OBs) of AcMNPV-wt or AcMNPV-31, respectively. C. The liquefication of the larvae that died from AcMNPV-31 was delayed compared with that of the larvae that died from AcMNPV-wt.

The 3rd instar S. exigua larvae were fed with different dosages of AcMNPV-wt or AcMNPV-31 to compare their larval survival rates (Fig. 4B). The survival rate of the tested larvae decreased with increasing viral dosage. Almost no larvae survived the feeding of AcMNPV-wt at dosages of 500, 1000, and 5000 OBs/larva. Larvae fed with AcMNPV-wt at dosages of 10, 50, and 100 OBs/larva obtained a survival rate of 47.2%, 36.2%, and 33.3%, respectively; 75.4% larvae survived from the feeding of AcMNPV-wt with a dosage of 5 OBs/larva (Fig. 4B). A higher survival rate was detected with AcMNPV-31-fed larvae compared to AcMNPV-wt-fed larvae at the same dose, in which only the highest dosage (5000 OBs/larva) obtained a complete mortality. The lethal dose analysis suggested that the LD₅₀ value of AcMNPV-31 (20.2 OBs/ larva) and the LD₉₀ value of AcMNPV-31 (1.7 × 10⁶ OBs/larva) was significantly higher than that of AcMNPVwt (16.7 OBs/larva, 500.7 OBs/larva).

Further lethal time analysis showed that the LT₅₀ value of AcMNPV-31 (7.8 days) and the LT₉₀ value of AcMNPV-31 (9.8 days) were slightly higher but not significantly different from those of AcMNPV-wt. These results indicate that the insecticidal properties of AcMNPV-31 are deteriorated, which might be due to the malformed nucleocapsids and decreased BV production resulting from the 3h-31 insertion. Interestingly, the larvae that died from the AcMNPV-31 infection remained intact × d after dying (Fig. 4C). The liquefaction of the AcMNPVwt-infected cadavers was detected on the first day after dying.

The histopathological observation was performed with the slides generated from AcMNPV-wt- or AcMNPV-31-infected S. exigua larvae to identify the response of larval tissues to virus infection. The appearance and histopathology of healthy larvae were normal, in which the midgut was darkly stained as a circle in the middle surrounded by light-stained fat bodies, and the tissues were free of infection from any pathogens (Fig. 5A). The wild type AcMNPV-infected larval midgut was missing, which might be due to degradation by the viral infection, and the resulting gut fragments could not be fixed during slicing (Fig. 5A). In addition, the nuclei of the fat bodies in the AcMNPV-wt-infected larvae were enlarged, and the hypodermal cell layers attached to the cuticles were stained with dark purple. The midgut of the AcMNPV-31-infected larvae was still distinguishable, but the cells were loosely arranged as compared with the midgut cellular array of healthy larvae (Fig. 5A).

Figure 5.  Histopathology of third instar S. exigua larvae infected with AcMNPV-wt or AcMNPV-31. A Hematoxylin- Eosin stain of transverse S. exigua larval sections. Slides of untreated healthy larva and AcMNPV-wt or AcMNPV-31 infected larva (72 hpi) specimens were observed. MG, mid gut; Ms, muscle; Cu, cuticle; FB, fat body; Magnification, 50×. B Enlarged cuticle of five S. exigua larval individuals infected by AcMNPV-wt or AcMNPV-31.

Pathological changes in the fat bodies, such as rounded cells and enlarged nuclei, were found in the AcMNPV-31- infected larvae, but the degree of lesion was not as serious as in the AcMNPV-wt-infected larvae. Moreover, the cuticle of the AcMNPV-31-infected larvae was unlike the cuticle of larvae infected with AcMNPV-wt, which was not covered by a layer of enlarged hypodermal cells (Fig. 5A). The cuticle comparison of larvae infected with AcMNPVwt or AcMNPV-31 is illustrated in Fig. 5B, with five different slides produced by five individuals infected by either AcMNPVwt or AcMNPV-31. The monolayer cell-layered cuticle of the AcMNPV-31-infected larvae were evidently different from several enlarged loose cell layers attached to the cuticle of the AcMNPV-wt-infected larvae (Fig. 5B). The pathological differences in the cuticle between the AcMNPV-wt- and AcMNPV-31-infected larvae may be the source of delayed liquefaction in AcMNPV-31-infected dead larvae.

The delayed liquefaction of larval cadaver and relatively intact inner tissues suggest that the infection of AcMNPV-31 might result in the inhibition of chitinase and cathepsin activities in the host. Therefore, enzyme activity was detected, and the results are shown in Fig. 6. Slight differences were found in the chitinase activity among the healthy, AcMNPV-wt-infected, and AcMNPV-31-infected larvae within the first 4 dpi (Fig. 6A). The chitinase activity of larvae infected with AcMNPV-wt was significantly increased at 4 dpi, which was 4.2 times higher than the chitinase activity of healthy larvae. The chitinase activity of AcMNPV-wt-infected larvaeat 5 dpi was 7.3 times higher than the chitinase activity of healthy larvae.

Figure 6.  Analysis of chitinase and cathepsin L activity in various AcMNPV variant-infected S. exigua larvae. A Determination of the chitinase activity of S. exigua larvae infected with AcMNPV-wt or AcMNPV-31. Error bars represent standard errors. Lowercase letters indicate significant differences based on analysis of variance with SPSS 22.0 (α = 0.5). B Determination of the cathepsin L activity of S. exigua larvae infected with AcMNPV-wt or AcMNPV-31. Error bars represent standard errors. Lowercase letters indicate significant differences based on analysis of variance with SPSS 22.0 (α = 0.5). C Expressional analysis of chitinase in S. exigua larvae infected with AcMNPV-wt or AcMNPV-31. Specific polyclonal antiserum against S. exigua chitinase-7 (SeCHIT7) and -11 (SeCHIT11) as well as the polyclonal antiserum against S. exigua chitin binding domain protein (SeCBD) were used to detect the chitinase expression level of S. exigua larvae infected with AcMNPV-wt or AcMNPV-31. The polyclonal antiserum against P10 was employed to detect the expression level of P10 encoded by AcMNPV. The commercially obtained mouse monoclonal antibody against His-tag was used to detect the expression of 3H-31 fused with EGFP in AcMNPV-31- infected S. exigua larvae. The polyclonal antiserum against GAPDH was employed to detect the expression level of GAPDH (reference protein) of the virus-infected S. exigua larvae. Lanes 1-3, total protein samples prepared from three healthy S. exigua larval individuals; lane M, pre-stained protein marker; lanes 4-6, total protein samples prepared from three S. exigua larval individuals infected with AcMNPV-wt (96 hpi); lanes 4-6, total protein samples prepared from three S. exigua larval individuals infected with AcMNPV-31 (96 hpi).

Compared to AcMNPV-wt infected larvae, the infection of AcMNPV-31 significantly inhibited the larval chitinase activity in the host, which decreased by 37.1% at 4 dpi (F = 31.68; d.f. = 2, 4; P = 0.0035) and 30.8% at 5 dpi (F = 31.68, d.f. = 2, 4; P < 0.0001). The cathepsin activity of healthy larvae was waved in a three-day period, which showed two high peaks at 2 dpi and 5 dpi (Fig. 6B). The increase in cathepsin activity in AcMNPV-wt-infected larvae began at 3 dpi, and the highest activity was observed at 5 dpi. In contrast, infection with AcMNPV-31 significantly inhibited cathepsin activity in the host compared with the AcMNPV-wt-infected larvae, which was only a quarter of the activity in AcMNPV-wt-infected larvae at 4 dpi (F = 348.5; d.f. = 2, 4; P < 0.0001) and one-fifth of that in AcMNPV-wt-infected larvae at 5 dpi (F = 348.5; d.f. = 2, 4; P < 0.0001).

The expression level of S. exigua chitinase (SeCHIT) in the AcMNPV-wt- or AcMNPV-31- infected larvae (96 hpi) was further confirmed by Western blotting analysis. The continuous detectable GAPDH in each repeat of the different larval protein samples provided a reference for the expressional analysis of SeCHITs (Fig. 6C). With the incubation of commercial His-tag monoclonal antibody, a specific ~ 28.2 kDa immunoreactive band (EGFP) was detected in the three samples produced from AcMNPV-wtinfected larvae, and an approximately 65.5 kDa immunoreactive band (3H-31 fused EGFP) was detected in the three samples produced from the AcMNPV-31-infected larvae, which indicated that the EGFP-fused 3H-31 was successfully expressed in the AcMNPV-31-infected larvae. A specific polyclonal antibody against P10 of AcMNPV was used to detect the infectious degree of the AcMNPV variants, and the bands that appeared in the AcMNPV-wtinfected larvae were more conspicuous than those in the AcMNPV-31-infected larvae, which might be due to the reduced BV production in AcMNPV-31-infected larvae. Two SeCHIT antibodies (anti-SeCHIT7 and antiSeCHIT11) were used to compare the expression levels of SeCHITs in healthy, AcMNPV-wt-infected, and AcMNPV-31-infected larvae (Fig. 6C). The results show that the expression levels of SeCHIT7 and SeCHIT11 in the AcMNPV-wt-infected larvae were evidently higher than those in the AcMNPV-31-infected larvae. A chitin binding domain-containing protein of S. exigua (SeCBD), which functioned as a helper of SeCHITs, had a similar expression level in both the AcMNPV-wt-infected and AcMNPV-31-infected larvae.

Unlike baculovirus infection, ascovirus infection commonly results in a non-liquefied cadaver. In order to clarify the variation pattern of the chitinase or cathepsin activities in HvAV-3h-infected larvae, temporal enzyme activity detection was performed with HvAV-3h-infected H. armigera or S. exigua larvae. As shown in Fig. 7A, chitinase activity was inhibited by the infection of HvAV-3h in both the H. armigera and S. exigua larvae. Inhibition of the cathepsin activity was also observed after infection with HvAV-3h, and the inhibition of cathepsin activity due to the HvAV-3h infection was more significant in the tested S. exigua larvae than in the tested H. armigera larvae.

Figure 7.  Determination of host larval enzyme activity changes after the RNAi of 3h-31. A Analysis of chitinase and cathepsin L activities in HvAV-3h-infected H. armigera or S. exigua larvae. B Expressional analysis of 3H-31 and chitinase in S. exigua larvae after the injection of 3h-31 or egfp dsRNA. The prepared antiserum against 3H-31 was used to detect the expression level of 3H-31 after the dsRNA injection; polyclonal antiserum against the MCP of HvAV-3h was used to detect the MCP expression level; GAPDH antiserum was used as the reference antibody. Lane M, pre-stained protein marker; lanes 1-3, total protein samples prepared from three S. exigua larval individuals infected with HvAV-3h (72 hpi); lanes 4-6, total protein samples prepared from three S. exigua larval individuals infected with HvAV-3h (72 hpi), followed by the injection of 1 μg 3h-31 dsRNA; lanes 7-9, total protein samples prepared from three S. exigua larval individuals infected with HvAV-3h (72 hpi), followed by the injection of 1 μg egfp dsRNA; lanes 10-12, total protein samples prepared from three S. exigua larval individuals infected with HvAV-3h (96 hpi). C Comparison of enzyme activities between the HvAV-3hinfected larvae and 3h-31 RNAi larvae. Third instar H. armigera and S. exigua larvae were used to perform the analysis. Error bars represent standard errors. Lowercase letters represent the significant differences between the different treatments (α = 0.05).

In order to detect the effects of 3H-31 on host larval chitinase and cathepsin activities, 3h-31 RNAi was performed on the HvAV-3h-infected H. armigera or S. exigua larvae. The synthesized dsRNA of 3h-31 and egfp was detected by agarose gel electrophoresis (shown in Supplementary Figure S5). Western blotting was performed to detect the expression level of 3H-31 after dsRNA injection in S. exigua larvae (Fig. 7B). Compared with egfp dsRNAinjected larvae, the 3h-31 dsRNA-injected larvae contained less 3H-31, which suggested that the RNAi of 3h-31 was succeeded. Compared with HvAV-3h-infected larvae (4 dpi), similar abundance of 3H-31 was observed in egfp dsRNA-injected larvae, which indicated that the injection of egfp dsRNA did not affect the expression of 3H-31 as well as the infection of HvAV-3h.

The MCP expression level of 3h-31 dsRNA-injected larvae was lower than that of the egfp dsRNA-injected larvae, which indicated that the silence of 3h-31 might also result in a decreased infection of HvAV-3h. Unsurprisingly, the larval cathepsin activity was significantly increased after the RNAi of 3h-31 compared to the egfp dsRNA-injected larvae in both the H. armigera (F = 28.45, d.f. = 12, P = 0.0478) and S. exigua (F = 84.55, d.f. = 12, P = 0.0422) (Fig. 7C). Contrary to our expected results, slight changes were detected in the larval chitinase activity, while no significant differences were detected between the chitinase activity in the 3h-31 dsRNA-injected and egfp dsRNA-injected larvae (F = 25.44, d.f. = 12, P = 0.9785 in the tested H. armigera; F = 32.14, d.f. = 12, P = 0.9987 in the tested S. exigua).

The main finding of the present study is that the 3h-31 gene is essential for HvAV-3h to control the activities of chitinase and cathepsin in the host larvae. First, 3H-31 was found to be a non-structural protein of HvAV-3h, which started to express at 3 hpi in the HvAV-3h-infected SeFB cells, but started to express at × hpi in the HvAV-3h-infected Ha-E cells (Fig. 1B). The differences in the transcription and expression phases of 3h-31 that were detected in the different cell lines might be due to the infectious tropism of ascovirus. Ascoviruses are tissue-specific infectious insect viruses with different specificities among different isolates. Spodoptera frugiperda ascovirus isolates are limited to the fat body, and H. virescens ascovirus isolates are reported to infect fat bodies, but also show mild infection of the epidermis and tracheal epithelium (Hamm et al. 1986, 1998; Federici et al. 1990; Yu et al. 2019a). A similar tropism was reported in HvAV-3e-infected H. zea cell lines, in which the fat body-derived cells (FB33) displayed a more rapid cytopathology than ovary-derived cells (HzAM1), although the MCP of HvAV-3e was expressed at 24 hpi in HvAV-3e-infected FB33 or HzAM1 cells (Asgari 2006).

The MCP of HvAV-3h was expressed at 12 hpi and accumulated over time in both infected Ha-E cells and SeFB cells (Fig. 1B). This suggested that the function of 3H-31 is completely different from that of MCP-viral proteins associated with DNA replication or inhibition of host immune reaction might have an earlier expression start point than that of viral structural protein, which is commonly started after viral DNA replication begins (Zaghloul et al. 2017, 2020). Another possible reason that caused the expression difference of 3H-31 in HvAV-3h-infected Ha-E or SeFB cells might be due to the host species selection of ascovirus. The in vivo differences in ascovirus-infected cells were reported by Yu et al. (2019b), in which the HvAV-3j-infected Spodoptera cells started to degrade 5 dpi, while the HvAV-3j-infected Helicoverpa cells remained intact even at 7 dpi. In any case, 3H-31 is a nonstructural protein of HvAV-3h, and it is transcribed and expressed before the transcription and expression of MCP.

Second, 3H-31 disrupted the BV production of AcMNPV, and the virulence of AcMNPV against S. exigua larvae decreased with the expression of 3H-31. Because of the lack of artificial bacterial chromosome of the ascovirus, the artificial bacterial chromosome of baculovirus (bMON14272) was employed to express 3H-31 in order to analyze its function. The constructed 3h-31-expressing bacmid (AcMNPV-31) had reduced BV production and viral DNA replication compared with those of the wildtype AcMNPV (Fig. 2DE). Furthermore, lucent tubes were found around the virogenic stroma of AcMNPV-31-infected SeFB cells (Fig. 3). A similar appearance of long tubular structures, which might be aberrant capsid structures, were reported in 38 k-deleted and vlf-1 deleted AcMNPV (Li et al. 2005; Vanarsdall et al. 2006; Wu et al. 2006). Both 38 K and VLF-1 are nucleocapsid proteins of AcMNPV, so their deletion would lead to the malformation of nucleocapsids (Yang and Miller 1998; Vanarsdall et al. 2004; Wu et al. 2008).

One of the differences in AcMNPV-31 was that normally shaped nucleocapsids can be found in the virogenic stroma, which means that the formation of nucleocapsids was not completely interrupted by the expression of 3H-31. The 3H-31 protein is a foreign protein of AcMNPV; it should not be a structural protein of the nucleocapsids of AcMNPV. The disrupted generation of nucleocapsids might be due to the DNA packaging into viral capsids. In this case, 3H-31, as a non-structural protein of HvAV-3h, might be associated with viral DNA replication and DNA transactions during viral DNA packaging. As the viral growth and viral replication were both inhibited by the expression of 3H-31, the virulence of AcMNPV-31 was decreased compared with the virulence of AcMNPV-wt.

Third, an interesting finding was that the liquefaction of AcMNPV-infected larvae was inhibited by the insertion of 3h-31 (Fig. 4C). Baculovirus can code chitinase (v-chiA) and cathepsin (v-CATH) to melt the host larval cadavers at the very late stage of infection, so as to release the viral progeny to commit their transmission between different host individuals. The baculovirus without v-chiA or v-CATH keeps the host cadavers intact after infection (Kang et al. 1998; Hom and Volkman 2002; Daimon et al. 2007; Hodgson et al. 2009; Vieira et al. 2012). Furthermore, v-chiA was revealed to be a proV-CATH folding chaperone that is involved in the maturation of v-CATH enzyme (Hom and Volkman 2000; Katsuma et al. 2009; Hodgson et al. 2011). The host cadavers failed in their liquefaction after the infection of AcMNPV-31, suggesting that the function of v-chiA and v-CATH was interrupted, and the following enzyme activity assays revealed that the total activity of chitinase and cathepsin L was inhibited by the infection of AcMNPV-31 compared with that of the larvae (Fig. 6). How 3H-31 manipulates the v-chiA or v-CATH encoded by the baculovirus and the chitinase or cathepsin of host larvae remain unknown.

Additionally, we demonstrated that the cathepsin activity of HvAV-3 h-infected H. armigera and S. exigua larvae was significantly increased after the silence of 3h-31 (Fig. 7). It is understandable that the ascovirus can inhibit the host larval growth to gain enough time and space to accomplish their replication and assembly. It has been reported that ascovirus infection can prolong larval lifespan for several weeks (Li et al. 2013; Hu et al. 2016; Chen et al. 2020). This phenomenon was in agreement with our investigation, which showed that the host larval chitinase and cathepsin were inhibited after infection with HvAV-3h (Fig. 7A), inhibiting the melting process of host larvae. It has been suggested that insect chitinase plays an important role in a variety of physiological processes, including growth, molting, and metamorphosis (Fukamizo 2000; Merzendorfer and Zimoch 2003; Wu et al. 2013), and might also have guidance and assistance functions on the activity of larval cathepsin. The increased host cathepsin activity but not the chitinase activity after the RNAi of 3h-31 suggested that 3H-31 might be associated with the maturation processes of cathepsin, but the speculation needs further verification.

The activity of host chitinase and cathepsin was inhibited by the appearance of 3H-31, which would keep the ascovirus-infected larvae remaining in their larval stages to facilitate the generation of viral progeny and transmission by parasitoid wasps. This mechanism is also practicable in the baculovirus infection. With the participation of 3H-31, the liquefaction of the host cadaver failed. These findings improve our understanding of the interaction between HvAV-3h and its host larvae (H. armigera and S. exigua).

We thank that Dr. Qilian Qin (Institute of Zoology, Chinese Academy of Sciences) provided the SeFB cells (IOZCAS-SpexII-A), Prof. Zhihong Hu (Wuhan Institute of Virology, Chinese Academy of Sciences) provided the Escherichia coli strain DH10B and Prof. Jianhong Li (College of Plant Science & Technology, Huazhong Agricultural University) provided Ha-E cells. This research was supported by the National Natural Science Foundation of China (32070168, 31700141, 31872027) and Provincial Natural Science Foundation of Hunan (2019JJ50234), and Changsha Science and Technology Project (kq1901033), and Double first-class construction project of Hunan Agricultural University, and Sakura Science Plan of Japan Science Technology Agency (JST).

HY and GHH conceptualized and designed the study. HY, YYOY, C-JY and NL performed the experiments in the study. HY and GHH contributed reagents to the study. HY and YYOY analyzed the data. HY, MN and GHH wrote the manuscript. All authors read and approved the final manuscript.

The authors declare that they have no conflict of interest.

The authors declare that this study is not involved in animal and human rights.