Two host microRNAs influence WSSV replication via STAT gene regulation

MicroRNAs (miRNAs) have important roles in post-transcriptional regulation of gene expression. During viral infection, viruses utilize hosts to enhance their replication by altering cellular miRNAs. The Janus kinase (JAK)/signal transducer and activator of transcription (STAT) pathway plays crucial roles in the antiviral responses. In this study, two miRNAs (miR-9041 and miR-9850) from Macrobrachium rosenbergii were found to promote white spot syndrome virus (WSSV) replication. The up-regulation of miR-9041 or miR-9850 suppresses STAT expression in the gills of M. rosenbergii, which subsequently down-regulates the expression of its downstream dynamin (Dnm) genes: Dnm1, Dnm2, and Dnm3. Knockdown of miR-9041 and miR-9850 restricts WSSV replication by up-regulating STAT and Dnm gene expression. The silencing of STAT, Dnm1, Dnm2, or Dnm3 led to an increase of the number of WSSV copies in shrimp. The injection of recombinant Dnm1, Dnm2, or Dnm3 proteins could inhibit WSSV replication in vivo. Overall, our research indicates the roles of host miRNAs in the enhancement of WSSV replication by regulating the host JAK/STAT pathway.

immunity against a broad range of viruses 23 . In addition to its involvement in virus endocytosis, Dnm has also been proposed to participate in membrane fusion between viruses and endosomes after endocytosis 24 .
MicroRNAs (miRNAs) have roles in the post-transcriptional regulation of gene expression 25 and various biological processes, such as proliferation, cell differentiation, apoptosis, tumorigenesis, and immune defense 25,26 . The production of mature miRNAs requires the participation of several molecules 27,28 . Typically, targeted mRNA leads to translation repression and/or mRNA degradation 29 . More miRNAs have been shown to participate in innate and adaptive immune response during virus infection by regulating the viral or host gene expression 30 . In humans, the Epstein-Barr virus (EBV)-encoded viral miR-BART22 modulates the viral gene product expression of the EBV latent membrane protein 2A (LMP2A), which may facilitate nasopharyngeal carcinoma carcinogenesis by evading the host immune response 31 . A human herpes virus miR-K12-11 attenuates IFN signaling and contributes to maintenance of viral latency by targeting I-κ -B kinase epsilon (IKKɛ ) 32 . In Marsupenaeus japonicus shrimp, the viral miRNAs WSSV-miR-66 and WSSV-miR-68 could target WSSV genes and further promote WSSV infection 33 . However, the host miRNA-mediated regulations of the STAT gene and its downstream genes have not been well studied in the giant freshwater prawn to date.
In this study, we demonstrated that miR-9041 and miR-9850 played positive roles in WSSV replication. The up-regulation of miR-9041 or miR-9850 suppresses STAT expression in the gills of M. rosenbergii, which subsequently down-regulates the expression of its downstream Dnm genes. The RNA interference (RNAi) of Dnm genes or the overexpression of Dnms by injection of recombinant proteins could enhance or inhibit virus replication, respectively. Our research describes the roles of host miRNAs in enhancing WSSV replication by regulating the host JAK/STAT pathway.

Effects of miR-9041 and miR-9850 on virus infection in shrimp. These 2 microRNAs (miR-9041
and miR-9850) was inducibly expressed in the WSSV challenged group in relative to the normal group (without WSSV challenge) based on the small RNA high-throughput sequencing data. So, we selected these 2 microRNAs for functional study. To further elucidate the roles of the host miR-9041 and miR-9850 in virus infection, both miRNAs were overexpressed in shrimp. When miR-9041 or miR-9850 was overexpressed in shrimp, the number of WSSV copies was examined. As shown in Fig. 1A, the overexpression of miR-9041 significantly increased the number of WSSV copies from 24 h to 48 h post-infection compared with the controls (miR-9041-scrambled and WSSV only). The overexpression of miR-9850 yielded similar results (Fig. 1C). By contrast, when miR-9041 or miR-9850 expression was knocked down by sequence-specific AMO-miR-9041 and AMO-miR-9850 in vivo, respectively, the WSSV copies were significantly decreased as compared with the controls (AMO-miR-9041scrambled, AMO-miR-9850-scrambled, and WSSV only; Fig. 1B,D). These findings indicated that the host miR-NAs play a positive role in virus replication.
Interaction between the host miRNAs and the host STAT gene. Increasing evidence has indicated that host miRNAs has important roles in host-virus interactions. To reveal the pathways mediated by host miR-NAs, the target genes of miR-9041 and miR-9850 were analyzed. The predictions by the TargetScan, miRanda, and Pictar algorithms showed that miR-9041 and miR-9850 could target the host STAT gene ( Fig. 2A,D), which is a gene involved in the antiviral immunity of shrimp 19 . Therefore, these host miRNAs may have significant effects on WSSV infection in shrimp.
To evaluate the interaction between the host miR-9041 or miR-9850 and the host STAT gene, we constructed the EGFP-STAT plasmid, which contained EGFP and the 3′-UTR of STAT in M. rosenbergii shrimp. The synthesized miR-9041 or miR-9850 mimic and the plasmid EGFP-STAT were co-transfected into insect High Five cells (Fig. 2B,E). The results showed that the fluorescence intensity in the co-transfected cells was significantly reduced compared with the intensity in the EGFP-STAT-transfected cells (Fig. 2C,F), thereby indicating that the synthesized miR-9041 or miR-9850 mimic repressed the expression of the STAT gene by targeting its 3′-UTR. However, the control miRNA mimic had a negligible effect on the expression of EGFP-STAT.
To further explore the interaction between miR-9041 or miR-9850 and STAT in vivo, the expression of miR-9041 and miR-9850 was overexpressed or silenced in shrimp, followed by the detection of STAT mRNA levels. The results indicated that miR-9041 or miR-9850 overexpression significantly decreased the STAT expression compared with the positive control (WSSV only), whereas the control miRNAs had no effect on the expression of STAT, thereby showing that miR-9041 and miR-9850 inhibited STAT expression in shrimp (Fig. 3A,C). When the expression of miR-9041 or miR-9850 was knocked down by AMO-miR-9041 or AMO-miR-9850, the expression of STAT was significantly up-regulated (Fig. 3B,D). These findings showed that miR-9041 and miR-9850 could interact with the host STAT gene in vivo.

Role of host STAT in virus infection.
The abovementioned data showed that the host STAT gene is the target of miR-9041 and miR-9850. Therefore, the role of STAT in virus infection was explored in shrimp. Results of qRT-PCR indicated that the STAT mRNA was detected in all the examined tissues and mainly expressed in the heart and intestine tissues of shrimp (Fig. 4A). In response to WSSV infection, the expression level of STAT was up-regulated from 24 h to 48 h, which was significantly higher than that of the untreated controls (Fig. 4B). The results suggested that the STAT gene may play an important role in virus infection.
To evaluate the influence of STAT in WSSV infection, the STAT expression was silenced by gene-specific siRNA (STAT-siRNA) in shrimp in vivo, followed by the detection of virus copies. The expression of the STAT gene was significantly knocked down at 24 and 48 h post-infection compared with the WSSV-only control group; by contrast, STAT-siRNA-scrambled had no effect on the STAT expression (Fig. 4C), thereby showing that the siRNA was highly specific. When the expression of the STAT gene was knocked down, the number of WSSV Scientific RepoRts | 6:23643 | DOI: 10.1038/srep23643 copies in shrimp gills was significantly increased compared with the control (WSSV only; Fig. 4D). The data indicated that the host STAT gene plays a negative role in WSSV infection in vivo.  To evaluate the roles of Dnm1, Dnm2, and Dnm3 in viral infection, these genes were characterized in shrimp. As shown in Figs. S1A, S1C, and S1E, the mRNAs of Dnm1, Dnm2, and Dnm3 were detected in all the examined tissues, including the hemocyte, heart, hepatopancreas, gills, stomach, intestine, and nerve tissues. In response to WSSV infection, the expression levels of Dnm1, Dnm2, and Dnm3 were significantly up-regulated from 24 h to 48 h (Figs. S1B, S1D, and S1F). Furthermore, when the expression of Dnm1, Dnm2, and Dnm3 was knocked down by sequence-specific siRNAs ( Fig. 7A-C), the number of WSSV copies significantly increased as compared with the controls (WSSV only). By contrast, the Dnm1-siRNA-scrambled, Dnm2-siRNA-scrambled, and Dnm3-siRNA-scrambled treatments had no effect on the virus infection ( Fig. 7D-F). The data demonstrated that Dnm1, Dnm2, and Dnm3 had significant effects on virus infection in shrimp. Recombinant plasmids (pET30a-Dnm1, pET30a-Dnm2, and pET30a-Dnm3) were respectively transformed into E. coli BL21 (DE3). After IPTG induction for 4 h, the whole cell lysates were analyzed by SDS-PAGE. In Fig. 8A, a distinct band with a molecular weight (MW) of approximately 100 kDa was detected, which was roughly consistent with the predicted MW of Dnm1 (theoretical MW, 96.4 kDa) and the ~5 kDa His tag at the N-terminal. The recombinant Dnm2 contained an additional N-terminal His Tag (5 kDa); thus, its size was relatively larger (~90 kDa) than the theoretical MW (84.8 kDa; Fig. 8B). Furthermore, a distinct band with an 80 kDa molecular mass was revealed, which agreed with the predicted molecular mass of the recombinant Dnm3 protein (77.6 kDa) with a His tag (Fig. 8C).

Dnm1, Dnm2, and
Given that Dnm1, Dnm2, and Dnm3 could be involved in virus infection at the mRNA level, we further hypothesized whether or not recombined Dnm1, Dnm2, and Dnm3 proteins affect the WSSV copies in shrimp. To test this hypothesis, the inhibition of WSSV replication was performed. The number of WSSV copies was replicated at a slower rate in the gills of the rDnm1, rDnm2, or rDnm3 injection groups compared with the BSA control and WSSV only groups at all post-injection times (24 and 48 h; Fig. 8D-F). These findings indicated that rDnm1, rDnm2, and rDnm3 had important functions in the inhibition of WSSV replication.

Discussion
MicroRNAs play an important role in the regulation of gene expression 25 . First discovered in the nematode Caenorhabditis elegans, miRNAs have been identified in all multicellular eukaryotes and some viruses 34 . Recent studies demonstrate that miRNAs can also strongly affect the replication of pathogenic viruses. Viruses are known to encode their own miRNAs and/or trigger changes in cellular miRNA expression to target mRNAs of virus and/or host genes, which further results in either mRNA degradation or translational repression 33 . The viral miRNAs WSSV-miR-66 and WSSV-miR-68 can promote WSSV infection by targeting the WSSV genes (the WSSV wsv094 and wsv177 genes are the targets of WSSV-miR-66; the wsv248 and wsv309 genes are the targets of WSSV-miR-68) 33 . WSSV-miR-N24 is employed by WSSV to regulate the expression of shrimp caspase 8 and facilitate viral replication by inhibiting the host antiviral apoptotic activity 35 . Cellular miRNAs can influence hepatitis B virus (HBV) replication directly by binding to HBV transcripts and indirectly by targeting cellular factors relevant to the HBV life cycle 36 . In human hepatocytes, the host miR-373 is up-regulated by the hepatitis C virus during its infection and negatively regulated by the type I IFN signaling pathway via suppression of JAK1 and IFN regulating factor 9 (IRF9) 37 . The host miR-146 targets the interleukin 1 receptor-associated kinase (IRAK1) and TNF receptor-associated factor 6 (TRAF6) as a negative regulator of constitutive nuclear factor-κ B (NF-κ B) activity in breast cancer 38 . However, the regulation of transcription factor expression by the simultaneous mediation of two different host miRNAs has not been explored to date. Our study highlights a novel aspect of two host To date, more than 10,000 miRNAs have been annotated in 96 species 39 , including over 2500 human miRNAs (miRBase ver. 21; http://www.mirbase.org/, released in June 2014). Each miRNA can regulate hundreds of different mRNAs, whereas a single mRNA can be conversely targeted by several miRNAs 40 . Cellular miRNAs play crucial roles in several biological processes, such as the innate and adaptive immune response; the deregulation of miRNA expression and function is also involved in various diseases 41 . Viruses can alter cellular miRNA expression; cellular miRNAs can positively or negatively influence virus replication during viral infection by regulating the expression of viral genes and/or cellular factors relevant to the course of virus-induced disease 42 . Viruses are equipped with complex machinery to exploit and manipulate the host pathways to establish an environment favorable for their persistence 43 .
In this study, shrimp miR-9041 and miR-9850 can regulate the expression of the host gene STAT, which is one of key components of the JAK/STAT pathway. A WSSV-encoded miRNA (WSSV-miR-22) could also promote WSSV infection in Marsupenaeus japonicas shrimp by targeting the host STAT gene 44 . The JAK-STAT signaling pathway plays a critical role in the initiation of antiviral response. An IFN-like antiviral cytokine known as Vago can activate the JAK/STAT pathway to control viral loads in West Nile virus-infected Culex quinquefasciatus cells 17 . The up-regulation of miR-9041 or miR-9850 suppresses the expression of STAT and its downstream molecules to promote WSSV replication. However, the downstream genes and their precise mechanisms are not yet fully understood.
In our previous study, thioester-containing protein TEP1 and TEP2 were the effectors of JAK-STAT signaling pathway 44 . Whereas in this study, Dnm genes were found to be downstream of STAT. Dnms are involved in The influence of STAT silencing on virus infection. The WSSV copies in gills of STAT-siRNA-treated shrimp were quantified using quantitative real-time PCR. In all panels, statistically significant differences between treatments were indicated by asterisks (**p < 0.01). antiviral immune defense. The Mx dynamin-like GTPases are key antiviral effector proteins of the types I and III IFN systems 23 . Human Mx2 or MxB are ISGs that contribute to the inhibition of the human immunodeficiency virus (HIV-1) replication by interferons 45 . A Dnm-dependent protein-trafficking pathway can mediate Ebola virus glycoprotein toxicity 46 . Dnm is required for recombinant adeno-associated virus type 2 (rAAV-2) infections, whereas the overexpression of mutant DnmI significantly inhibited AAV-2 internalization and gene delivery 47 . Therefore, Dnm is a key mediator involved in a broad range of viral infections.
In conclusion, we demonstrated that the host miR-9041 and miR-9850 play positive roles in WSSV replication by targeting the host JAK/STAT signal pathway. Dnm genes are regulated by STAT and are involved in antiviral immune response. Therefore, WSSV infection could induce host miRNAs, which inhibit the JAK/STAT pathway and finally enhance virus replication.
For each treatment, 20 shrimp were used. At different times post-infection (0, 24, 36, and 48 h), 5 shrimp were randomly collected for each treatment and subjected to the subsequent analyses. All the experiments were biologically replicated three times.

Prediction of target genes.
To predict the genes targeted by host miRNAs, the shrimp genome sequence (data unpublished) was employed with three independent computational algorithms TargetScan 5.1 (http://www. targetscan.org), miRanda (http://www.microrna.org/), and Pictar (http://pictar.mdc-berlin.de/). TargetScan was used to search for miRNA seed matches (nucleotides 2-8 from the 5′-end of miRNA) in the 3′-untranslated region (UTR) sequences. miRanda was used to match the entire miRNA sequences. The miRanda parameters were set as free energy < − 20 kcal/mol and score > 50. Pictar was employed to search the combined effects of microRNA target microRNA-based or other characteristics, with score > 20. Finally, the results predicted by the three algorithms were combined, and the overlaps were calculated 33,49 . Plasmid construction. To characterize the direct interaction between miR-9041 or miR-9850 and the shrimp STAT gene, the 3′-UTR of STAT and the enhanced green fluorescent protein (EGFP) gene were cloned into a pIZ/EGFP V5-His vector (Invitrogen, USA). The EGFP gene was amplified from the pEGFP vector (BD Biosciences, USA) with EGFP-specific primers (5′-AAGAGCTCGGATCCCCGGGTA-3′ and 5′-AATCTAGAGTCGCGGCCGCTTTA-3′). Subsequently, the STAT 3′ -UTR was cloned into the pIZ vector downstream of the EGFP gene with the XbaI and SacII restriction sites and the sequence-specific primers (5′-GCTCTAGATAATAGGTTCTAGCACATG-3′ and 5′-TCCCCGCGGATGTATATTATAAAAGTTTC-3′). As controls, the STAT 3′-UTR sequence (GTGAATT) complementary to the miR-9041 seed sequence was mutated to TGTCCGG, thereby yielding the EGFP-∆ STAT-9041 construct, whereas the STAT 3′ -UTR sequence  The sequence-specific Dnm1-siRNA, Dnm2-siRNA or Dnm2-siRNA was injected into shrimp. Then the gene expression was examined. As a negative control, siRNAscrambled was included in the injections. WSSV alone was used as a positive control. The detection of WSSV copies after Dnm1 (D), Dnm2 (E) or Dnm3 (F) knocked down in shrimp. The gills of siRNA-treated shrimp were subjected to the quantitative real-time PCR analysis to examine the WSSV copies. WSSV alone was used as a positive control. In all panels, asterisks indicated significant differences between treatments (**p < 0.01). the cells reached 70-80% confluence, they were transfected with 6 μg of EGFP, EGFP-STAT, EGFP-∆ STAT-9041, or EGFP-∆ STAT-9850. Simultaneously, the cells were transfected with 300 nM of synthesized miR-9041, synthesized miR-9850, or synthesized control miRNA. All the miRNAs were synthesized by Shanghai Gene Pharma Co., Ltd. (Shanghai, China). The transfection reactions were performed in triplicate with Cellfectin transfection reagent (Invitrogen) according to the manufacturer's protocol. After 12 h of incubation, the transfected cells were seeded into 96-well plates at a concentration of 2.0 × 10 4 cells per well. At 48 h after transfection, the fluorescence of the cells was examined with a Flex Station II microplate reader (Molecular Devices, USA) at 490/510 nm for excitation and emission (Ex/Em), respectively. The fluorescence values were corrected by subtracting the autofluorescence of cells that did not express EGFP. All the experiments were biologically replicated three times.
The RNAi assay was conducted in shrimp by the injection of a siRNA at 30 μg/shrimp into the lateral area of the fourth abdominal segment with a 1 mL sterile syringe. The siRNA (15 μg) and WSSV (10 5 copies/mL) were co-injected into virus-free shrimp at a volume of 100 μL/shrimp. At 16 h after co-injection, the siRNA (15 μg; 100 μL/shrimp) was injected into the same shrimp. Simultaneously, the siRNAs-scrambled (15 μg; 100 μL/shrimp) was co-injected into virus-free shrimp. At 16 h after co-injection, siRNAs-scrambled (15 μg; 100 μL/shrimp) was injected into the same shrimp. WSSV (10 5 copies/mL; 100 μL/shrimp) alone was injected as a positive control. As a negative control, PBS was injected instead of the siRNAs. For each treatment, 20 shrimp were used. Shrimp gills were collected at different times after the last injection (0, 24, 36, and 48 h). From each treatment, 5 shrimp specimens were randomly selected and collected for later use. The assays were biologically replicated three times.