Abstract

   MicroRNAs (miRNAs), are a novel class of gene expression regulators,
   that have been found to participate in regulating host–virus
   interactions. However, the function of insect-derived miRNAs in
   response to virus infection is poorly understood. We analyzed miRNA
   expression profiles in the fat bodies of Helicoverpa armigera (H.
   armigera) infected with Mamestra brassicae multiple
   nucleopolyhedroviruses (MbMNPV). A total of 52 differentially expressed
   miRNAs (DEmiRNAs) were filtered out through RNA-seq analysis. The
   targets of 52 DEmiRNAs were predicted and 100 miRNA–mRNA interaction
   pairs were obtained. The predicted targets of DEmiRNAs were mainly
   enriched in the Wnt signaling pathway, phagosome, and mTOR signaling
   pathway, which are related to the virus infection. Real-time PCR was
   used to verify the RNA sequencing results. ame-miR-317-3p, mse-miR-34,
   novel1-star, and sfr-miR-6094-5p were shown to be involved in the host
   response to MbMNPV infection. Results suggest that sfr-miR-6094-5p can
   negatively regulate the expression of four host genes eIF3-S7, CG7583,
   CG16901, and btf314, and inhibited MbMNPV infection significantly.
   Further studies showed that RNAi-mediated knockdown of eIF3-S7
   inhibited the MbMNPV infection. These findings suggest that
   sfr-miR-6094-5p inhibits MbMNPV infection by negatively regulating the
   expression of eIF3-S7. This study provides new insights into MbMNPV and
   H. armigera interaction mechanisms.

   Keywords: microRNA, MbMNPV, Helicoverpa armigera, RNA-seq,
   sfr-miR-6094-5p, eIF3-S7, host–virus interaction

1. Introduction

   Helicoverpa armigera (H. armigera) is a worldwide pest [[44]1] that
   damages several crops, such as maize, wheat, cotton, pepper, and
   tomato, and causes heavy losses [[45]2]. Baculoviridae family comprises
   arthropod-specific viruses with double-stranded DNA genomes of 80–180
   kb [[46]3]. Mamestra brassicae multiple nucleopolyhedrovirus (MbMNPV)
   is a member of baculoviridae family, which is characterized by a wide
   host spectrum and known to be effective in the field control of H.
   armigera infection. Bioassay shows that MbMNPV has great application
   and has been developed as a commercial bio-pesticide in China [[47]4].
   However, not much is known about the molecular mechanism underlying the
   MbMNPV–host interactions.

   RNA interference (RNAi) is the process by which double-stranded RNA
   specifically silences the expression of its homologous genes through
   the degradation of their cognate mRNA [[48]5]. RNAi has shown enormous
   potential to be developed as a novel pest control strategy [[49]6].
   MicroRNAs (miRNAs) are a class of small non-coding RNAs that have
   proven roles in the regulation of gene expression by binding to the
   transcripts of targets, which can be pivotal in disease and pest
   control as well as virus–insect interactions [[50]7]. Studies have
   found that the overexpression of Dicer-2 and Ago-2 genes can improve
   the cell defense ability against the acute infection of the cricket
   paralysis virus (CrPV) [[51]8]. Another study reports Ecdysis of
   Nilaparvata lugens fed with nlu-miR-173 agomirs was compromised, which
   affected their normal growth [[52]9]. Another report shows miR-14
   positively regulates the H. armigera ecdysone receptor (Ha-EcR) during
   single nucleopolyhedrovirus (HaSNPV) infection. Consistent with
   decreased levels of miR-14 following HaSNPV infection, the transcript
   levels of its target gene Ha-EcR declined, which indicates that miR-14
   is involved in the regulation of H. armigera growth and is related to
   virus–host interactions [[53]10]. In conclusion, miRNAs play key roles
   in the growth of insects and offer an approach for controlling insect
   pests by inhibiting the growth and development of the pests [[54]11].

   The fat body of insects is a vital organ for biosynthesis, metabolism,
   detoxification, and control of insect ecdysis. It has become a
   significant research model for the study of metabolic disorders
   [[55]12]. In this study, the changes in miRNA levels in the fat body of
   H. armigera in response to MbMNPV infection were investigated by the
   use of high-throughput sequencing. Our results proved insights into the
   role of miRNAs in H. armigera–MbMNPV interactions, which help us to
   further understand host–baculovirus interactions and provide new
   strategies for insect pest control by miRNA.

2. Materials and Methods

2.1. Insects and Viruses

   The eggs of H. armigera were purchased from Jiyuan Baiyun Industry Co.,
   Ltd. (Jiyuan, China). Insects were fed with an artificial diet and
   reared in the laboratory at 27 ± 2 °C, 60 ± 2% relative humidity with a
   photoperiod of 16:8 (light:dark) [[56]13].

   MbMNPV was provided by our laboratory and later propagated by feeding
   to third instar H. armigera. Occlusion bodies (OBs) of MbMNPV were
   purified as described by Kyei-Poku and Kunimi [[57]14]. The OBs were
   purified from MbMNPV-infected cadavers by differential centrifugation
   followed by sucrose density gradient ultracentrifugation.

2.2. Bioassay and Sample Collection

   The newly molted fourth-instar larvae with weights ranging from 0.32 g
   to 0.37 g, were fed with an artificial diet contaminated with MbMNPV
   occlusion bodies (OBs; 1 × 10^8 OBs/larva) after starving for 20 h.
   Upon consumption of the diets, the larvae were fed with fresh diets
   without exposure to any pathogens. Mock-infected larvae were fed with
   the same quantity of diets, but coated with distilled water.
   Twenty-five larvae per group were evaluated for larval weight at 24,
   48, and 72 h after infection with MbMNPV. The fat body was collected at
   72 h after infection with MbMNPV. Larvae were anesthetized on ice for
   30 min, the upper epidermis of the larvae was cut off, and the fat body
   was collected into a 1.5 mL Eppendorf tube on ice.

2.3. Identification of H. armigera Infected with MbMNPV

   The copy number of MbMNPV was calculated by absolute quantification. In
   short, MbMNPV genomic DNA was extracted from the liquefied larvae
   infected with MbMNPV and used as the template for the amplification of
   the conserved polyhedrin gene by PCR. The PCR products were sub-cloned
   into the pMD™19-T Vector Cloning Kit (TaKaRa). The recombinant plasmid
   polyhedrin-pMD-19T was used to calculate the plasmid copy number and
   generate a standard curve for RT-qPCR. CT value was obtained by RT-qPCR
   using the specific primer of polyhedrin for the fat body of mock- and
   MbMNPV-infected larvae for 72 h, and then the copy number of MbMNPV was
   determined by standard curve. The β-actin gene was used as an
   endogenous reference. The primers were synthesized by Sangon Biotech
   Co., Ltd. (Shanghai, China) ([58]Table S1).

2.4. RNA Extraction and Quality Control

   Mock-infected and MbMNPV-infected larvae consisted of three repeats
   labeled as CK-1, CK-2, and CK-3 and MbMNPV-1, MbMNPV-2, and MbMNPV-3,
   respectively. At 72 h post-infection (hpi), total RNA was extracted
   using the mirVana miRNA Isolation Kit (Ambion) according to the
   manufacturer’s protocol. The RNA quality and purity were evaluated
   using agarose gel electrophoresis, Nanodrop microvolume
   spectrophotometer, and Agilent 2100 Bioanalyzer.

2.5. Library Preparation and Sequencing

   Small RNA sequencing libraries were generated using the NEBNext Small
   RNA Library Prep Set for Illumina Kit (Cat. No. NEB#E7330S, NEB, USA)
   following the manufacturer’s recommendations. Briefly, 1 μg of total
   RNA was ligated with the 3′ adaptor and 5′ adaptor at each end. After
   reverse transcription, PCR amplification was performed and the products
   with fragments ranging from 140 to 160 bp in length were gel purified.
   Library quality was assessed on the Agilent Bioanalyzer 2100 system
   using DNA high-sensitivity chips. The library preparations were finally
   sequenced using the Illumina Novaseq 6000 platform.

2.6. MiRNAs Sequencing and Analysis

   Sequencing data were analyzed by removing adaptor dimers, common RNA
   families (rRNA, scRNA, Cis-reg, snRNA, tRNA), and the sequences with
   repeats and low complexity by using Bowtie [[59]15]. Subsequently, the
   sequences (15–41 nt) were mapped to the H. armigera genomic database
   against the miRBase v22 database [[60]16] ([61]http://www.mirbase.org/,
   August 2021) to identify known miRNAs and novel 3p- and 5p-derived
   miRNAs. The novel miRNAs based on their precursors were identified
   using the MiRDeep2 [[62]17]. Clean data were imported into the miRDeep2
   software, and the novel miRNAs were finally obtained by entering the
   running command (mapper.pl input.fa -c -j -l 18 -m -p genome.fa -s
   reads_collapsed.fa -t reads_collapsed_vs_genome.arf -v and miRDeep2.pl
   reads_collapsed.fa genome.fa reads_collapsed_vs_genome.arf miRNA.fa
   mature_other_miRNA.fa hairpin.fa)

   The expression level of miRNAs was normalized by TPM (transcript per
   million) method. The differentially expressed miRNAs (DEmiRNAs) were
   selected with the thresholds of q-value < 0.05 and |log2 (foldchange)|
   > 1. The miRanda [[63]18] and RNAhybrid software tools were used to
   predict the target of DEmiRNAs by comparing the candidate miRNA with
   the 3′-untranslated region (UTR) sequence of mRNAs. GO enrichment and
   KEGG pathway enrichment analysis of DEmiRNA targets were respectively
   performed using R based on the hypergeometric distribution. The top 30
   GO terms and the top 20 of KEGG enrichment were considered to be
   enriched by targets.

2.7. Construction of DEmiRNA–mRNA Regulatory Network

   The targets of DEmiRNAs were predicted using miRanda, RNAhybrid, and
   PITA software tools. A total of 100 mRNAs were predicted by taking the
   intersection of targets predicted by these three software tools and
   miRNA–mRNA interaction pairs were obtained. To identify the potential
   miRNA–mRNA interaction relationship, the Cytoscape 3.7.2 software was
   used to visualize the miRNA–mRNA regulatory network by importing
   miRNA–mRNA interaction pairs data into the Cytoscape software, which
   shows the inverse correlation of the miRNA and target expression.

2.8. Validation of miRNA and mRNA by Real-Time Quantitative PCR

   To validate the RNA-seq data, three miRNAs were randomly selected for
   RT-qPCR confirmation with ChamQ™ universal SYBR^® RT-qPCR master mix
   (Vazyme, China). Total RNAs were extracted using RNA Isolater Total RNA
   Extraction Reagent (Vazyme, China). For miRNA determination, 1 µg of
   total RNA was reversed using the miRNA 1st Strand cDNA Synthesis Kit
   (by stem-loop) (Vazyme, China) with miRNA-specific stem-loop primers.
   For gene quantification, the cDNA was prepared using HiScript^® III
   All-in-one RT SuperMix Perfect for RT-qPCR (Vazyme, China) with the
   gDNA wiper. The PCR was conducted with a StepOnePlus™ Real-Time PCR
   System (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA)
   under the following conditions: 95 °C for 30 s, followed by 40 cycles
   of 95 °C for 10 s and 60 °C for 30 s. The specificity of each primer
   pair was ensured by analyzing the melting curve. Experiments were
   performed with three independent biological replicates and technical
   replicates. The relative quantity of miRNAs and mRNAs was analyzed by
   RT-qPCR and β-actin was used as the reference gene. All primers were
   synthesized by the Sangon Biotech Co., Ltd. (Shanghai, China)
   ([64]Table S1). Data were analyzed by the 2^−ΔΔCT method.

2.9. Synthesis of dsRNA, RNA Interference and Viral Infection

   Synthesis and microinjections of dsRNA were performed as previously
   described [[65]19]. Briefly, target fragments of GFP (green fluorescent
   protein) and HaeIF3-S7 for dsRNA synthesis were amplified using
   specific primers. The PCR products were sub-cloned into the 5 min
   TA/Blunt-Zero Cloning Kit (Vazyme, China) and used as templates for
   target sequence amplification. By using a T7 RNAi Transcription Kit
   (Vazyme, China), the dsRNAs were synthesized in vitro with specific
   primers fusing the T7 promoter at the 5′ end and dissolved in DEPC
   water. The purity and yield of the dsRNA were checked on a 1.0% agarose
   gel and a spectrophotometer, respectively. The primers used to
   synthesize dsRNA are shown in [66]Table S2.

   In the knockdown experiment, newly molted fifth-instar larva was
   incubated for 30 min on ice, then injected with 10 μg single dsRNA
   solution (ds GFP and ds HaeIF3-S7) using a 10 µL microsyringe
   (Hamilton) in their last leg. Besides, an equivalent volume of DEPC was
   injected for the control.

   The newly molted fifth-instar larvae injected with dsRNA solution were
   fed with fresh artificial diets for 4 h. The above larvae, after
   overnight starvation, were then fed with artificial diets coated with
   an MbMNPV suspension (OBs; 1 × 10^7 OBs/larva). Upon consumption of the
   diets, the larvae were fed with fresh diets without exposure to any
   pathogens. Subsequently, three larvae infected with mock and MbMNPV
   were randomly selected at 24 h and 48 h, and the tissues of the fat
   body were dissected. To investigate the RNAi efficiency, RT-qPCR
   analysis was performed for the control group. To reveal the impact of
   RNAi HaeIF3-S7 on MbMNPV infection, the relative expression of
   polyhedrin was performed by RT-qPCR for the mock-infected and
   MbMNPV-infected groups.

2.10. Data Analysis

   All the data in this study are presented as the means ± SEM. The
   statistical significance was determined by Student’s t-test for
   unpaired comparisons between two different groups, and p < 0.05 was
   regarded as statistically significant. All statistical analyses were
   performed with Prism software.

3. Results

3.1. MbMNPV Infection Caused Significant Reduction in Larval Wet Weight

   The wet weight of the mock-infected and MbMNPV-infected larvae was
   evaluated. Compared with the mock-infected group, the wet weight of the
   MbMNPV-infected larvae had a significant reduction ([67]Figure 1). With
   the extension of infection time, despite the increase in larval weight,
   the wet weight was still significantly lower than that of the
   mock-infected group. This indicated that MbMNPV infection inhibited the
   growth of H. armigera.

Figure 1.

   [68]Figure 1
   [69]Open in a new tab

   Effect of MbMNPV infection on body weight of H. armigera. ns, p > 0.05;
   ***, p < 0.001; p < 0.05 indicated the significant difference
   (Student’s t-test). Each experiment was performed in three replicates,
   and data are shown as mean ± s.e.m.

3.2. Verification of the MbMNPV-Infected H. armigera

   To verify the MbMNPV infection, the fat bodies of mock and
   MbMNPV-infected larvae were extracted and the infection was verified by
   PCR amplification of the polyhedrin gene. Our results indicated that
   the H. armigera were infected with MbMNPV successfully, while the
   control group was not infected with the virus ([70]Figure 2A,B). The
   positive samples were used for further analysis.

Figure 2.

   [71]Figure 2
   [72]Open in a new tab

   Verification for infection of H. armigera with MbMNPV. (A) PCR
   verification of polyhedrin gene of mock-infected and MbMNPV-infected
   larvae. Note: 1~6 are the amplification of polyhedrin gene in CK-1,
   CK-2, and CK-3 and MbMNPV-1, MbMNPV-2, MbMNPV-3. (B) The copy number of
   MbMNPV in the mock-infected and MbMNPV-infected larvae fat body. *** p
   < 0.001.

3.3. MiRNA Sequencing and Data Analysis

   Six small RNA libraries (mock-infected and MbMNPV-infected larvae,
   three repeats labeled as CK-1, CK-2, and CK-3, and MbMNPV-1, MbMNPV-2,
   and MbMNPV-3, respectively) were sequenced and the quality of the
   sequencing data was analyzed ([73]Table 1). Clean reads were
   distributed at 23.28 M~24.12 M, with Q30 of 95.94~96.98%. A total of
   775 miRNAs were identified, including 616 known miRNAs and 159 novel
   miRNAs ([74]Figure 3). The length distribution of the known miRNA in
   all samples is shown in [75]Figure 4. The results showed that most of
   the miRNA reads were 22 nt in size.

Table 1.

   Quality evaluation of sample sequencing data.
   Sample CK_1 CK_2 CK_3 MbMNPV_1 MbMNPV_2 MbMNPV_3
   Raw reads 29.55 M 29.17 M 31.7 M 27.16 M 27.2 M 27.34 M
   Sequences with Q20 Greater than 80% 23.92 M 24.13 M 23.32 M 23.83 M
   23.97 M 24.16 M
   Clean reads 23.88 M 24.08 M 23.28 M 23.79 M 23.93 M 24.12 M
   Q30 95.94% 96.84% 96.71% 96.31% 96.79% 96.98%
   GC content 55.31% 56.90% 56.94% 57.62% 57.20% 59.02%
   [76]Open in a new tab

Figure 3.

   [77]Figure 3
   [78]Open in a new tab

   MiRNAs distribution in samples. (A) The distribution of total miRNA in
   the two samples. (B) The distribution of known miRNA in the two
   samples. (C) The distribution of novel predicted miRNA in the samples.

Figure 4.

   [79]Figure 4
   [80]Open in a new tab

   Length distribution of known miRNA.

3.4. Functional Analysis for Targets of DEmiRNAs

   The DEmiRNAs between the MbMNPV-infected group and the control were
   screened based on TPM values (q < 0.05 and |log2 (foldchange)| > 1). A
   total of 52 miRNAs were differentially expressed, including 24
   upregulated and 28 downregulated miRNAs ([81]Figure 5A). The volcanic
   and heat maps ([82]Figure 5B) of the normalized expression of miRNAs
   were generated and the expression patterns of the DEmiRNAs in the
   infection process were analyzed.

Figure 5.

   [83]Figure 5
   [84]Open in a new tab

   Screening and enrichment analysis of DEmiRNAs in MbMNPV-infected group
   compared with control group. (A) The volcano map of different miRNA.
   (B) Top 30 significantly enriched GO terms of predicted targets for
   DEmiRNAs in molecular function, cellular components, and biological
   process. (C) Hierarchical clustering analysis (heatmap) for DEmiRNAs.
   (D) Top 20 significantly enriched KEGG analyses of predicted targets
   for DEmiRNAs.

   GO and KEGG analyses for the targets of DEmiRNAs were performed to
   investigate the function of DEmiRNAs ([85]Figure 5C,D). GO analysis
   indicated that the predicted targets were mainly enriched in the
   following terms: positive regulation of transcription by RNA polymerase
   II (GO:0045944), ubiquitin-dependent protein catabolic process
   (GO:0006511), regulation of transcription by RNA polymerase II
   (GO:0006357), under biological process category; nucleus (GO:0005634),
   cytoplasm (GO:0005737), integral component of membrane (GO:0016021),
   under cellular component category; ATP binding (GO:0005524), DNA
   binding (GO:0003677), Rab GTPase binding (GO:0017137), under molecular
   function category ([86]Figure 5C). In the KEGG pathway analysis, the
   Wnt signaling pathway (ID: haw04310), phagosome (ID: haw04145), mTOR
   signaling pathway (ID: haw04150), and cysteine and methionine
   metabolism (ID: haw00270) were significantly enriched ([87]Figure 5D)
   ([88]Table S3).

3.5. MiRNA Target Prediction and miRNA–mRNA Interaction Network

   The targets of 52 DEmiRNAs were predicted using miRanda, RNAhybrid, and
   PITA software. A total of 100 mRNAs were predicted by taking the
   intersection of targets predicted by these three software tools, which
   were associated with 15 DEmiRNAs, including 8 upregulated and 7
   downregulated miRNAs. The inverse correlation of the miRNAs and target
   expression was analyzed and visualized by the Cytoscape software
   ([89]Figure 6). Our analysis showed that except for novel4_star,
   novel5_star, sfr-miR-2755-5p, and bmo-miR-2755-3p for which only one
   target was predicted, other miRNAs could regulate multiple mRNAs. Four
   miRNAs (ame-miR-317-3p, mse-miR-34, sfr-miR-6094-5p, novel1-star) were
   predicted to have 9, 10, 44, and 4 mRNA targets, respectively.

Figure 6.

   [90]Figure 6
   [91]Open in a new tab

   Interaction networks of differentially expressed miRNAs and putative
   target mRNAs in fat body of MbMNPV-infected H. armigera. The square
   nodes and circular nodes represent miRNAs and mRNA, respectively.
   Downregulated and upregulated RNAs are indicated by green and red
   colors, respectively.

3.6. Verification of DEmiRNAs by Real-Time Quantitative PCR

   Three miRNAs (pca-bantam-3p, bmo-miR-2755-3p, and cqu-miR-10-5p) were
   randomly selected from the DEmiRNAs to verify the sequencing data by
   RT-qPCR. The results showed that the expression of these three miRNAs
   in the MbMNPV-infected group was significantly upregulated compared
   with the control group ([92]Figure 7). The RT-qPCR results were
   consistent with that of the sequencing results, indicating that the
   sequencing data was reliable.

Figure 7.

   [93]Figure 7
   [94]Open in a new tab

   Relative expression of miRNAs in control and MbMNPV. ***, p < 0.001; p
   < 0.05 indicate significant difference (Student’s t-test). Each
   experiment was performed in three replicates, and data are shown as
   mean ± s.e.m.

3.7. Relative Expression of miRNAs in H. armigera Fat Body

   Four DEmiRNAs (ame-miR-317-3p, mse-miR-34, sfr-miR-6094-5p,
   novel1-star) were selected based on the GO and KEGG analysis. Our
   sequencing results showed that the expression of ame-miR-317-3p,
   sfr-miR-6094-5p, and mse-miR-34 was downregulated and that of
   novel1-star was upregulated, at 72 h post-infection.

   The relative expression of four miRNAs, at 24, 48, 60, and 72 h, in the
   virus infection group and control group was analyzed by RT-qPCR
   ([95]Figure 8). The results showed that the relative expression of
   miR-317-3p in virus-infected fat bodies of H. armigera was low at 24 h.
   However, with the extension of virus infection time, its relative
   expression level gradually increased. The relative expression of
   miR-317-3p increased up to 60 hpi, which was different from the
   sequencing data. It may be caused by the host individual differences
   and sampling time ([96]Figure 8A); the relative expression of
   mse-miR-34 was 4.8 times and 3.2 times higher at 48 h and 60 h in
   comparison to mock control, respectively. However, its expression at 72
   h decreased in comparison to the control group ([97]Figure 8B). The
   relative expression of miRNA novel1-star was significantly lower at 24
   h, which increased at 48 h, 60 h, and 72 h ([98]Figure 8C) in
   comparison to the mock control. The expression of sfr-miR-6094-5p was
   significantly upregulated at 24 h, and gradually the expression
   declined at 48 h and 60 h and was significantly downregulated at 72 h
   in comparison to the mock control ([99]Figure 8D). The relative
   expression of mse-miR-34, novel1-star, and sfr-miR-6094-5p was
   consistent with the sequencing results.

Figure 8.

   [100]Figure 8
   [101]Open in a new tab

   Relative expression of miRNAs at different time points in the control
   and MbMNPV-infected group. ns, p > 0.05; *, p < 0.05; **, p < 0.01;
   ***, p < 0.001, p < 0.05 indicate significant difference (Student’s
   t-test). Each experiment was performed in three replicates, and data
   are shown as mean ± s.e.m.

   The results proved that the miRNA expression profile of the host was
   indeed changed after virus infection at different treatment times,
   which indicated that these miRNAs may be involved in H. armigera–MbMNPV
   interaction. It is necessary to screen miRNA targets for follow-up
   research.

3.8. DEmiRNAs Inhibited Expression of Predicted Target Genes

   By using the bioinformatics software, miRanda and RNAhybrid, the
   targets of sfr-miR-6094-5p were predicted. Based on GO and KEGG
   analysis, four miRNA–mRNA interaction pairs, eIF3-S7 (LOC110371347:
   [102]XM_021327532.1), CG7583 (LOC110380909: [103]XM_021341066.1),
   CG16901 (LOC110381560: [104]XM_021341895.1) and btf314 (LOC110382609:
   [105]XM_021343259.1), were selected for further analysis.

   To verify the relationship of sfr-miR-6094-5p and predicted targets
   eIF3-S7, CG7583, CG16901, and btf314, overexpression and inhibition
   assay of sfr-miR-6094-5p were performed in vivo. The sfr-miR-6094-5p
   agomir or antagomir and their relative controls were injected into H.
   armigera larvae (fifth-instar, 1st day). Sequences for miRNA agomir and
   antagomir are shown in [106]Table S4. The results are shown in
   [107]Figure 9. Agomir injection of sfr-miR-6094-5p decreased the
   expression of eIF3-S7, CG7583, CG16901, and btf314 at 48 h
   post-treatment, while antagomir injection significantly increased these
   variables. In conclusion, the injection of sfr-miR-6094-5p agomir and
   antigomir could negatively regulate the expression of host gene
   eIF3-S7, CG7583, CG16901, and btf314 at 48 h, of which eIF3-S7 and
   CG7583 have been reported to be involved in host–virus interactions,
   indicating that sfr-miR-6094-5p may play a vital role in regulating
   host–virus interactions.

Figure 9.

   [108]Figure 9
   [109]Open in a new tab

   The relative expression of eIF3-S7 (A), CG7583 (B), CG16901 (C) and
   btf314 (D) after injection of sfr-miR-6094-5p agomir and antagomir. ns,
   p > 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001; p < 0.05 indicate
   significant difference (Student’s t-test). Each experiment was
   performed in three replicates, and data are shown as mean ± s.e.m.

3.9. The Inhibition of sfr-miR-6094-5p on MbMNPV Infection

   To verify the impact of sfr-miR-6094-5p on MbMNPV infection, the
   sfr-miR-6094-5p agomir, antagomir, and their controls were injected
   into the MbMNPV-infected larvae, respectively. The relative expression
   of MbMNPV polyhedrin gene at 24 h post-infection, was detected by
   RT-qPCR. As shown in [110]Figure 10, after injecting sfr-miR-6094-5p
   agomir, the expression of the polyhedrin gene in the experimental group
   was significantly downregulated and it was upregulated after antagomir
   treatment. These results indicated that sfr-miR-6094-5p negatively
   regulated MbMNPV infection in H. armigera larvae.

Figure 10.

   [111]Figure 10
   [112]Open in a new tab

   The relative expression of MbMNPV polyhedrin gene after injection of
   sfr-miR-6094-5p agomir and antagomir in the MbMNPV-infected larvae. *,
   p < 0.05; **, p < 0.01; p < 0.05 indicate significant difference
   (Student’s t-test). Each experiment was performed in three replicates,
   and data are shown as mean ± s.e.m.

3.10. Knockdown of HaeIF3-S7 Inhibited the MbMNPV Infection

   As sfr-miR-6094-5p negatively regulates the target gene and MbMNPV
   infection, the effect of RNAi HaeIF3-S7 on MbMNPV infection was further
   investigated. After successful knockdown of eIF3-S7 ([113]Figure 11A),
   we observed a significant decrease in MbMNPV replication ([114]Figure
   11B). Interestingly, at 48 h post-infection, the wet weight of the
   larvae with HaeIF3-S7 knocked down was downregulated ([115]Figure
   11C,D). In conclusion, our results show that the knockdown of eIF3-S7
   inhibited MbMNPV infection.

Figure 11.

   [116]Figure 11
   [117]Open in a new tab

   The effects of HaeIF3-S7 RNAi on the larvae. At 1d fifth-instar, larvae
   were injected with dsRNA of HaeIF3-S7 (ds eIF3-S7), dsRNA of green
   fluorescent protein (dsGFP), and DEPC (control). (A) The expression
   levels of the eIF3-S7 gene. (B) The expression levels of the polyhedrin
   gene. (C,D) The MbMNPV-infected larvae weight were analyzed. The
   standard error is represented by the error bar. ns, p > 0.05; *, p <
   0.05; **, p < 0.01; ***, p < 0.001; p < 0.05 indicate significant
   difference (Student’s t-test). Each experiment was performed in three
   replicates, and data are shown as mean ± s.e.m.

4. Discussion

   MiRNAs play pivotal roles in insect–baculovirus interactions [[118]20].
   Therefore, in the current study, we performed miRNA sequencing in the
   fat body of MbMNPV-infected H. armigera. A total of 52 DEmiRNAs were
   obtained, of which 24 were upregulated and 28 were downregulated. Of
   these miRNAs, bmo-miR-1175-3p [[119]21], miR-12 [[120]22], miR-92b
   [[121]23], miR-34 [[122]24], miR-927 [[123]25], miR-281 [[124]26],
   miR-927 [[125]25], and miR-317 [[126]27] have been reported in previous
   studies and known to be related to virus infection. Reports suggest
   that one specific miRNA can regulate the expression of multiple mRNAs
   and vice versa [[127]28]. In the current study, the interaction
   networks of 8 upregulated and 7 downregulated miRNAs with their mRNA
   targets revealed similar patterns. The KEGG pathway enrichment analysis
   revealed that the predicted targets of DEmiRNAs were mainly enriched in
   the following signaling pathways: Wnt signaling pathway [[128]29], mTOR
   signaling pathway [[129]30], phagosome [[130]31], and cysteine and
   methionine metabolism [[131]32], which are related to virus infection.

   In this study, ame-miR-317-3p, mse-miR-34, novel1-star, and
   sfr-miR-6094-5p were identified as DEmiRNAs in MbMNPV-infected H.
   armigera. It has been reported that the miR-34 family participates in
   immune response and regulates virus replication [[132]24]. Among
   patients with liver fibrosis and hepatocellular carcinoma infected
   because of hepatitis B virus (HBV) infection, miR-34a-5p inhibits liver
   fibrosis by regulating the TGF-β1/Smad3 pathway [[133]33]. It has been
   confirmed that miR-317 can be used as a negative regulator of toll
   immune response which affects Drosophila survival [[134]34]. Since
   novel1-star is similar to pca-bandam-3p in clustering ([135]Figure 5B),
   it may have functions similar to pca-bandam-3p.

   In the current study, sfr-miR-6094-5p negatively regulates the
   expression of eIF3-S7, CG7583, CG16901, btf314, and against MbMNPV.
   eIF3-S7 and CG7583 have been reported to be involved in virus
   interactions. The eukaryotic initiation factor 3 (eIF3), a
   multi-subunit complex composed of 13 subunits (a-m), participates in
   translation initiation, termination, and ribosomal recycling [[136]35],
   which may be the reason for the wet weight reduction in larvae. It has
   been reported that eIF3 is associated with virus replication. For
   example, overexpression of eIF3 can affect the replication of the
   yellow fever virus [[137]36] and facilitates the replication of swine
   fever virus NS5A and the production of virus particles in offspring
   [[138]37], which is consistent with our observation that the knockdown
   of Ha eIF3-S7 inhibited MbMNPV replication. Our results showed that
   sfr-miR-6094-5p inhibited MbMNPV infection by negatively regulating the
   expression of eIF3-S7.

   The C-terminal binding protein CG7583 (CTBP) is enriched in Wnt
   signaling and Notch signaling pathways (haw04310, haw04330). The Wnt
   signaling pathway is involved in many biological processes [[139]38].
   It has been found that virus replication can be delayed by interfering
   with the interaction of CTIP and CTBP with miRNA [[140]39]. Marek’s
   disease virus (MDV) encodes a major oncoprotein Meq, which is related
   to the interaction of CTBP [[141]40]. CTBP also interacts with two
   oncoproteins, EBNA3A and EBNA3C, which are encoded by the Epstein Barr
   virus nuclear antigen 3C virus [[142]41]. CG16901, as an RNA binding
   protein, is enriched in the Dorso-ventral axis formation signal pathway
   (haw04320). Zhang identified 17 differentially expressed proteins in
   studying the proteomics of the lymphoid organs of vibrio anguillarum
   shrimp, including a similar one to the squid CG16901-PC protein. The
   host gene btf314 is homologous to the transcription factor eIF3 and is
   enriched in the apoptosis fly pathway (haw04214), which plays a crucial
   role in maintaining the normal functioning of the body. Our results
   indicated that sfr-miR-6094-5p may also inhibit the MbMNPV infection by
   negatively regulating the expression of host CG7583, CG16901, and
   btf314, which needs to be further validated.

   In conclusion, we profiled the differentially expressed miRNAs in the
   fat body of MbMNPV-infected H. armigera at 72 hpi and analyzed their
   interaction networks and putative functions. sfr-miR-6094-5p inhibits
   MbMNPV infection by negatively regulating the expression of eIF3-S7 in
   vivo. In the future, we will analyze the molecular mechanism of
   expression change of sfr-miR-6094-5p induced by MbMNPV, miR-6094-5p
   inhibiting the expression of eIF3 in H. armigera. This research can
   contribute a theoretical reference for insect pest control and shed
   light on further understanding of host–virus interactions.

Acknowledgments