A MicroRNA-1–Mediated Inhibition of the NF-κB Pathway by the JAK-STAT Pathway in the Invertebrate Litopenaeus vannamei

The JAK-STAT and NF-κB pathways are independent signaling cascades that both are conserved across vertebrate and invertebrate species and play critical regulatory roles in various biological processes (1, 2). The NF-κB family has five members in mammals (RelA, RelB, c-Rel, p105/p50, p100/p52), three in Drosophila (Dorsal, Relish, and Dif) and two in crustaceans (Dorsal and Relish) (3, 4). The activation process of the canonical NF-κB pathway in response to pathogen infection consists of multiple cascade steps, including recognition of invading pathogens by TLRs, intracellular recruitment of MyD88 and IL receptor–associated kinases (IRAKs), formation of the MyD88/IRAKs/TNF receptor–associated factor 6 (TRAF6) complex, activation of IκB kinases, degradation of IκB in the IκB/NF-κB complex, releasing and nuclear translocation of the transcription factor NF-κB, and activation of downstream effector genes (5, 6). In contrast, there are seven STATs in mammals (STATs 1–4, STATs 5a/5b, and STAT6) and one in arthropods, all of which exert functions in the two-intracellular–component JAK-STAT pathway (7, 8). Activation and nuclear translocation of STAT occurs upon JAK-mediated tyrosine phosphorylation triggered by ligand binding-mediated cell receptor clustering (9–11). Notably, the JAK-STAT pathway constitutes the downstream part of the IFN system, which is central to antiviral defense in mammals (12, 13). After binding to receptors, IFNs activate the JAK-STAT pathway to induce expression of hundreds of IFN-stimulated genes, leading to the establishment of the cellular antiviral state (14, 15).

Because of the lack of acquired immunity, invertebrates rely solely on innate immune system to defend against infectious agents, and thus providing several ideal platforms for studies of innate immunity and cellular signaling mechanisms (16–18). The Pacific white shrimp Litopenaeus vannamei, a representative species of the subphylum Crustacea, is currently the most cultured shrimp worldwide. Because of its important evolutionary status and the need for disease control in aquaculture, the immune system of L. vannamei has attracted more and more research attention. L. vannamei has a moderate body size and is more suitable for artificial infection than other invertebrate species, making it a useful model for studying invertebrate immunity (19). Studies in L. vannamei have proved fruitful in systematically characterizing the molecular architecture of the canonical TLR–NF-κB pathway, which consists of multiple components that are homologous to those of Drosophila and vertebrates, such as TLRs, MyD88, Tube (the homolog of IRAK4), Pelle (IRAK1), TRAF6, Cactus (IκB), and Dorsal (a NF-κB family member) (4, 20–25). Similar to that of Drosophila, the JAK-STAT pathway of L. vannamei is a simple and effective model with one JAK and one STAT and has also been extensively studied (26, 27). These provide a basis for further elucidating the regulatory mechanism of invertebrate immune system using L. vannamei as a research platform. Interestingly, studies in shrimp have suggested that both TLR–NF-κB and JAK-STAT signaling pathways are implicated in antiviral and antibacterial responses (28–31), indicating a possible overlap of the immune mechanisms in which they are involved. However, little is known about the relationship and interaction between these two pathways in invertebrates so far.

MicroRNAs (miRNAs) are conserved endogenous small noncoding RNAs (18–26 nt) derived from primary transcripts (pri-miRNAs, ∼1000 nt) and stem-loop–shaped precursors (70–100 nt) (32, 33). Through incorporating into a RNA-induced silencer complex–loading complex in an ATP-dependent manner, the mature miRNA interacts with the partially complementary binding site in the 3′-untranslated region (3′ UTR) or open reading frame of the target mRNA to inhibit protein translation and/or to modulate mRNA stability (34–36). Many miRNAs have been shown to target various signaling components of JAK-STAT or NF-κB pathways in invertebrates. For instance, in Drosophila, miR-279 targets the JAK/STAT ligand Upd to control rest/activity rhythms (37). miR-9041 and miR-9850 from Macrobrachium rosenbergii suppress STAT expression to attenuate the antiviral immune response (38). In L. vannamei, miR-1959, whose expression is regulated by Dorsal, can promote the activation of Dorsal by targeting Cactus, and thus mediating a positive feedback loop of the NF-κB pathway to regulate immune response (39). These finding reveal the important roles of miRNAs in invertebrate intracellular signaling transduction. In the current study, based on elucidation of the transcriptional regulatory mechanism and identification of the target gene of the miRNA-1 (miR-1) molecule, we established a miRNA-mediated inhibition of the NF-κB pathway by the JAK-STAT pathway in L. vannamei. We demonstrated that this mechanism was involved in balancing the antiviral and antibacterial responses, which may help to further understand the cellular signaling system of invertebrates.

L. vannamei (∼10 g) were obtained from an aquaculture farm in Zhuhai, China. Five percent of shrimp were randomly sampled to ensure they were free of Vibrio parahaemolyticus and white spot syndrome virus (WSSV) by PCR as previously described (40, 41). Shrimp were acclimated at ∼28°C for at least 7 d in a recirculating water tank system filled with air-pumped seawater before experiments. Preparation of the stocks of V. parahaemolyticus and WSSV was performed as previously described (39).

The sequence of mature miR-1 was obtained from a small RNA–sequencing library and verified using stem-loop real-time RT-PCR and Northern blot, following methods as previously reported (39). The specific primers and probes used were shown in Supplemental Table I. The genome sequence of the primary miR-1 (pri–miR-1) gene and its promoter was retrieved from the reported L. vannamei complete genome using BLAST (42). The pri–miR-1 promoter was analyzed using Neural Network Promoter Prediction (http://www.fruitfly.org/seq_tools/promoter.html) and Jaspar (http://jaspar.genereg.net/). The precursor miR-1 (pre–miR-1) sequence was analyzed using RNA-folding form software (http://mfold.rna.albany.edu/?q=mfold/RNA-Folding-Form), amplified by RT-PCR using total RNA extracted from L. vannamei gill, and cloned into the pAc5.1/V5-His A vector (Invitrogen) with flanking sequences. The constructed vector was transfected into Drosophila S2 cells, and the expression of miR-1 was analyzed using stem-loop real-time RT-PCR and Northern blot.

Total RNA was extracted from shrimp tissues using an RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) and reverse-transcribed into cDNA with primer miR1-RT (Supplemental Table I) using a PrimeScript Reverse Transcription Kit (Takara, Shiga, Japan). Stem-loop real-time RT-PCR was performed on a LightCycle 480 System (Roche Diagnostics) at a final volume of 10 μl containing 1 μl of cDNA, 5 μl of 2 × SYBR premix ExTaqTM II (Takara), and 500 nM of each primer (miR1-qRTF/miR1-qRTR or U6-qRTF/U6-qRTR, Supplemental Table I). The thermal cycling parameters were 2 min at 95°C, followed by 40 cycles of 15 s at 95°C, 15 s at 60°C, and 10 s at 72°C. After the cycling protocol, melting curves were obtained by increasing the temperature from 72°C to 95°C (0.5°C/s). The expression levels of miR-1 were calculated using 2^−ΔΔCt method after normalization to U6 RNA.

Total RNA (10 μg) was electrophoretically separated on a 12% denaturing polyacrylamide gel and was transferred onto a positively charged nylon membrane (Roche Diagnostics) using a Trans-Blot SD Semi-Dry Transfer Cell System (Bio-Rad Laboratories). The membrane was cross-linked by UV (254 nm) and prehybridized at 68°C for 30 min using ULTRAhyb Ultrasensitive Hybridization Buffer (Ambion), and hybridized overnight with miR-1 and U6 LNA probes (Exiqon, Vedbaek, Denmark, Supplemental Table I) at a final concentration of 0.1 nM at 65°C, respectively. The membrane was then washed, blocked, and incubated with anti–digoxigenin-AP Ab (Roche Diagnostics). The signals were detected using CDP-Star Chemiluminescent Substrate (Roche Diagnostics) and developed on x-ray films (Fujifilm, Tokyo, Japan) with a short exposure time of 10 s.

The wild-type miR-1 promoter (pro–miR-1) and its STAT binding site–mutated (Mut) form generated by overlapping PCR were cloned into the pGL3-Basic Vector (Promega) to generate pGL3–pro-miR1 and pGL3-pro–miR-1–Mut, respectively. The expression vectors for STAT, JAK, and suppressor of cytokine signaling 2 (SOCS2) genes, pAc5.1-STAT, pAc5.1-JAK, and pAc5.1-SOCS2, were generated based on pAc5.1/V5-HisA vector (Invitrogen). The 3′ UTR of MyD88 was amplified by RT-PCR, and mutation of the miR-1 target site was achieved by overlapping PCR. The wild-type and mutated MyD88 3′ UTR were cloned into a previously reported vector pGL3-249 after the open-reading frame of luciferase to generate pGL3-MyD88 or pGL3-MyD88-Mut vectors, respectively (39). The primers used were summarized in Supplemental Table I.

Dual-luciferase reporter assays were performed following a reported method (39). Briefly, S2 cells were cultured on a 96-well plate with 70% confluent at 28°C in Schneider insect medium (Sigma-Aldrich) supplemented with 10% FBS (Life Technologies). To analyze the activity of miR-1 promoter, the protein expression vector or the control pAc5.1/V5-GFP (100 ng for each well) was cotransfected with pGL3-pro–miR-1 or pGL3-pro–miR-1–Mut (50 ng for each well) using FuGENE HD transfection reagent (Promega). For miRNA target identification, 100 nM of chemosynthesized miR-1 or negative control (NC) mimic (GenePharma, Shanghai, China) was cotransfected with 50 ng of pGL3-MyD88 or pGL3-MyD88-Mut into S2 cells. The pRL-TK (50 ng for each well) (Promega) Renilla luciferase plasmid was used as an internal control. Each treatment was repeated for eight wells. At 48 h posttransfection, the activity of luciferase was detected using a dual-luciferase reporter assay system (Promega).

The pAc5.1-STAT or pAc5.1-GFP plasmid was transfected into S2 cells, and 48 h later, cells were collected and the nuclear proteins were extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific). The 5′ biotin-labeled probe derived from miR-1 promoter (−1145 to −1067 bp) that contains the STAT binding site was synthesized by Invitrogen. The unlabeled or the STAT site–deleted probes were used as control. EMSA was performed using a LightShift Chemiluminescent EMSA Kit (Thermo Fisher Scientific) as previously described (39). A rabbit Ab against shrimp STAT (GL Biochem, Shanghai, China) was used for supershift assays.

For tissue distribution analysis of miR-1, tissues were sampled and pooled from nine healthy L. vannamei. For challenge experiments, each shrimp was i.m. injected with 50 μl of PBS buffer containing 10⁶ copies of WSSV or 10 μg of poly(I:C). Shrimp injected with PBS buffer containing no stimulants were used as the control. The gill and hemocytes were sampled from nine shrimp at each time point of 0, 4, 12, 24, 48, 72, and 96 h postinjection and subjected to total RNA purification, stem-loop real-time RT-PCR, and Northern blot to detect the expression of miR-1.

To investigate the regulatory mechanism of miR-1, dsRNAs specific to STAT, JAK, SOCS2, and GFP (as control) genes were synthesized using a T7 RiboMax express RNAi System (Promega), following the method as previously described (43), and i.m. injected into shrimp at a dose of 10 μg (diluted in 50 μl of PBS). At 48 h postinjection, shrimp were stimulated with V. parahaemolyticus (10⁵ CFUs), and 4 h later, hemocytes and gill were sampled. Total nuclear protein of hemocytes (each sample from 30 shrimp) was extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific) and detected by Western blot using a rabbit anti-shrimp STAT Ab (GL Biochem). Expression of miR-1 in gill and hemocytes (each sample pooled from nine shrimp) was analyzed using stem-loop real-time RT-PCR and Northern blot, as described above.

The 5′-end cholesterol–modified miRNA mimic agonist miR-1 (agomiR-1) and inhibitor antagonist miR-1 (antagomiR-1) and their controls (agomiR-NC and antagomiR-NC, Supplemental Table I) were synthesized by GenePharma. To analyze the effect of miR-1 on expression of immune related genes, healthy L. vannamei were injected with 50 μl of PBS-diluted 10-μg miRNA mimic or inhibitor, and 48 h later, gill and hemocytes were sampled from each group with each sample pooled from nine shrimp. Expression of antimicrobial peptides (ABPs) penaeidins (PEN2, PEN3, and PEN4) and anti-LPS factors (ALF1, ALF3, and ALF-AVK), C-type lectins (CTL1, CTL2, CLT3, and CTL4), and lysozymes (Lys-IT and Lys), were detected using real-time RT-PCR with specific primers (Supplemental Table I).

To elucidate the activation of the Dorsal pathway, shrimp were injected with 10 μg of miR-1 mimic/inhibitor or STAT dsRNA (mixed or not mixed with 10 μg of miR-1 mimic) and, 48 h later, were stimulated with 5 μg of LPS, followed by tissue sampling at 4 h poststimulation. For each sample, total protein of hemocytes and gill pooled from nine shrimp was extracted using T-PER Tissue Protein Extraction Reagent (Thermo Fisher Scientific). Total nuclear protein of hemocytes pooled from 30 shrimp was extracted and analyzed by Western blot using rabbit Abs against shrimp MyD88, Cactus, and Dorsal proteins (GL Biochem). The Abs against histone H3 (CST) and β-actin (Medical and Biological Laboratories, Aichi, Japan) were used to detect the nuclear and cytoplasmic internal control proteins, respectively. The gray values of the specific protein bands were calculated using Quantity One 4.6.2 software (Bio-Rad Laboratories) by Gauss model and normalized to those of internal control proteins.

Hemocytes were washed with 2× Leibovitz L-15 medium (Life Technologies) triply, stained with Dil, mixed with FITC-labled V. parahaemolyticus at a 1:100 ratio of cells/bacteria, and incubated at 28°C for 1 h. After three washes with L-15 medium, hemocytes were detected using flow cytometry, and the Dil and FITC double fluorescence signals were used to identify cells that have phagocytized bacteria. The thresholds and boundaries of fluorescence signals were set based on detection of the controls Dil-stained hemocytes and FITC-labeled V. parahaemolyticus phagocytized by unstained cells (Supplemental Fig. 1). A total of 500,000 events were detected for each sample.

Hemolymph smear samples were made on siliconized slides and fixed with 4% paraformaldehyde for 10 min. After treatment with 1% Triton X-100 for 20 min, slides were successively incubated with rabbit Ab against shrimp Dorsal (GL Biochem) together with mouse Ab against β-actin (Medical and Biological Laboratories) and Alexa Fluor 488–conjugated goat anti-rabbit Ab (Abcam) together with Alexa Fluor 594–conjugated goat anti-mouse Ab (CST). After staining with Hoechst 33342 (Invitrogen) for the nuclei, slides were observed using a Leica LSM 410 Confocal Microscope (Wetzlar, Germany).

Shrimp (n = 50 in each group) were injected with miR-1 mimic, inhibitor, or controls and, 48 h later, were challenged with WSSV (10⁶ copies) or V. parahaemolyticus (10⁵ CFUs). Experiments were done triply, and the cumulative mortality was recorded. Parallel challenge experiments were also set to sample gill and muscle. Total DNA extraction was performed using the DNeasy Blood and Tissue Kit (Qiagen). Specific primers for the VP28 (GenBank accession No. AY422228) and 16S rDNA (EU660325) genes were used to analyze the copies of WSSV and V. parahaemolyticus in tissues, respectively, by relative quantitative real-time PCR with the L. vannamei elongation factor 1-α (EF1-α) (GenBank accession No. GU136229) as internal control.

All experiments were replicated at least three times. Results of Western blot were analyzed using ImageJ software. Statistical comparisons were performed by two-tailed unpaired Student t test or one-way ANOVA followed by Dunnett post hoc test using SPSS 16.0 software. Data are mean ± SD. Data and considered significant at p < 0.05 and are indicated by *p < 0.05 and **p < 0.01.

The sequence of L. vannamei miR-1 retrieved from a high throughput small RNA sequencing library shares the same sequence with that of Drosophila and differs from those of human and mouse with 2 nt at positions 20 and 22 (Fig. 1A). The expression of miR-1 in L. vannamei tissues was analyzed using stem-loop real-time PCR and Northern blot, which gave consistent results (Fig. 1B, 1C). miR-1 was widely expressed in all the detected tissues, with high levels in eyestalk, heart, muscle, scape, and stomach and relatively low levels in other tissues. Besides mature miR-1 molecules, the Northern blot result also showed a clear band of pre–miR-1 (Fig. 1C). Based on the reported L. vannamei genome data (42), the sequence of the pre–miR-1 was retrieved and analyzed. The 76-bp pre–miR-1 was predicted to fold a hairpin structure with a minimal free energy of −36.90 kcal/mol and harbor the mature miR-1 in its 3′ stem (Fig. 1D). To verify the sequence of pre–miR-1, the putative pre–miR-1 was cloned into expressing vector and transfected into S2 cells (Fig. 1E). Stem-loop RT-PCR demonstrated that the level of miR-1 in pre–miR-1 vector–transfected cells was much higher than that in control cells that had been transfected with empty vectors and expressed low levels of endogenous miR-1 (Fig. 1F). Consistent with this, Northern blot showed specific bands of the mature form and precursor of miRNA in pre–miR-1 vector–transfected cells, the positions of which coincided with those of the endogenous miR-1 (Fig. 1G).

The transcriptional initiation site of pri–miR-1 was mapped at 350 bp upstream of pre–miR-1 by bioinformatics analysis (Fig. 2A, Supplemental Fig. 2). To verify the prediction and identify the promoter of miR-1, a 2273-bp-long sequence upstream the putative transcriptional initiation site was subjected to dual-luciferase reporter assays (Fig. 2B). Compared with the control, the analyzed sequence efficiently drived the transcription of luciferase, suggesting that it contained the miR-1 promoter. Bioinformatics analysis predicted that the miR-1 promoter contains a potential STAT binding motif (5′-ACGGTTTCTTGGGAA-3′) located in −249 to −235 (Fig. 2A). Dual-luciferase reporter assays showed that compared with the control, STAT/JAK alone- and combined-treatment increased the activity of the wild-type but not the STAT binding motif–mutated miR-1 promoter (Fig. 2B). The interaction between STAT and miR-1 promoter was determined using EMSA. The results demonstrated that STAT bound the biotin-labeled probe of miR-1 promoter to form a retarded shift band, which was further retarded to form a supershift band when anti-STAT Ab was added (Fig. 2C). The shift and supershift bands were eliminated when 200 × unlabeled probes were added to competitively bind STAT. In contrast, no shift or supershift bands of the labeled probes of STAT binding site–mutated miR-1 promoter were observed (Fig. 2D). These suggested that miR-1 could be directly regulated by STAT. To further explore the regulatory mechanism of miR-1 expression, several components of the JAK-STAT pathway was silenced using RNAi strategy in vivo (Fig. 2E). Consistent with our expectation, silencing of STAT and JAK inhibited the nuclear translocation of STAT upon V. parahaemolyticus stimulation, whereas silencing of SOCS2, the inhibitor of the JAK-STAT pathway, enhanced it (Fig. 2F). As a result, both stem-loop real-time PCR and Northern blot demonstrated that expression of miR-1 was decreased upon silencing of STAT and JAK but increased upon silencing of SOCS2 (Fig. 2G, 2H), confirming that miR-1 was a target gene of JAK-STAT pathway.

Activation of the JAK-STAT pathway is well known as one of the most important events in immune defense against viral infection (8) and can be induced by WSSV infection and exogenous nucleic acid stimulation in shrimp (26, 27). We then examined the expression profile of miR-1 in shrimp postinfection by WSSV and stimulation by the viral dsRNA mimic poly(I:C). Stem-loop real-time PCR demonstrated that the level of miR-1 in both gill and hemocytes did not change significantly in PBS mock-stimulated shrimp but increased in WSSV-infected or poly(I:C)–stimulated shrimp (Fig. 3A, 3B). The effect of WSSV infection on the expression of miR-1 was more obvious than that of poly(I:C)–stimulation. Similar trend of miR-1 expression in hemocytes was observed by Northern blot, which demonstrated that the levels of both mature miR-1 and pre–miR-1 were upregulated after immune stimulation (Fig. 3C). These confirmed that the expression of miR-1 could be induced by viral infection.

The regulatory target of miR-1 was analyzed by RNAhybrid software, which demonstrated that miR-1 was imperfectly complementary with the 3′ UTR of the Dorsal pathway component MyD88 (Fig. 4A). Dual-luciferase reporter assays showed that miR-1 decreased the expression of luciferase reporter gene with the wild-type but not the miR-1 seed binding site–mutated MyD88 3′ UTR (Fig. 4B). In vivo experiments also demonstrated that compared with the control (agomiR-NC), miR-1 mimic (agomiR-1) significantly decreased the protein level of MyD88 in hemocytes and gill, the two major immune tissues in shrimp, whereas miR-1 inhibitor (antagomiR-1) exerted an opposite effect on it (Fig. 4C, 4D).

Given that MyD88 is the upstream adaptor of the TLR-Dorsal pathway that directly transfers signals from TLR, we next explored the effect of miR-1 on LPS-induced activation of the TLR-Dorsal pathway, the hallmarks of which are degradation of the Doral inhibitor protein Cactus and nuclear translocation of Dorsal (20, 39). The results of Western blot demonstrated that compared with the control, the protein level of Cactus was increased in gill and hemocytes of miR-1 mimic–treated shrimp but decreased in those of miR-1 inhibitor–treated shrimp (Fig. 4E, 4F), suggesting that miR-1 inhibited the degradation of Cactus. Correspondingly, the level of nuclear-translocated Doral was decreased in hemocytes from miR-1 mimic–treated shrimp but increased in those from miR-1 inhibitor–treated shrimp (Fig. 4G). In vitro experiments on S2 cells also demonstrated that miR-1 attenuated the promoting effect of MyD88 on the nuclear translocation of GFP-tagged Dorsal induced by LPS (Supplemental Fig. 3). The results were further confirmed by immunofluorescence on hemocytes in vivo, which demonstrated that miR-1 mimic suppressed the nuclear translocation of Dorsal upon LPS stimulation, whereas miR-1 inhibitor promoted it (Fig. 4H). These suggested that miR-1 could inhibit the activation of TLR-Dorsal pathway through targeting MyD88.

Considering that STAT can directly regulate the expression of miR-1, we next explored the effect of STAT on activation of the Dorsal pathway. The in vivo expression of STAT and its upstream inhibitor SOCS2 was silenced using RNAi strategy, and shrimp were further stimulated with LPS to activate the TLR-Dorsal pathway. Compared with the control, upon the silencing of STAT, the protein level of MyD88 was increased (Fig. 5A) and that of Cactus was reduced (Fig. 5B), and the nuclear translocation of Dorsal was enhanced (Fig. 5C). In contrast, after the silencing of SOCS2, the degradation of Cactus was promoted (Fig. 5B), whereas the expression of MyD88 and nuclear translocation of Dorsal was attenuated (Fig. 5A, 5C). These suggested that STAT could negatively regulate the activation of the Dorsal pathway. To explore the role of miR-1 in the regulatory effect of STAT on the Dorsal pathway, exogenous miR-1 mimic was injected into the STAT-silenced shrimp to compensate for the decreased expression of endogenous miR-1. Compared with the control, miR-1 mimic attenuated the reducing effect of STAT silencing on the expression of MyD88 (Fig. 5D), degradation of Cactus (Fig. 5E), and nuclear translocation of Doral (Fig. 5F), indicating that miR-1 could be implicated in the regulation of the Dorsal pathway by STAT.

The above results demonstrated that miR-1 could function as a bridge connecting the two major immune regulatory pathways JAK-STAT and TLR-Dorsal. We next investigated the roles of miR-1 in immunity in vivo. The expression of a set of canonical immune effector genes in gill and hemocytes showed similar trends after miR-1 mimic or inhibitor injection. Compared with the control, miR-1 mimic downregulated the expression of ABPs PEN-2, -3, and -4, ALF-1, -3, and -AVK, C-type lectins CTL-1, -2, -3, and -4, and lysozymes Lys and Lys-IT, whereas the miR-1 inhibitor exhibited opposite effects, that is, enhancing the expression of them (Fig. 6A, 6B). Notably, the effect of miR-1 mimic or inhibitor on CTLs and Lys was obviously stronger than that on ABPs. The effect of miR-1 on the phagocytosis of hemocytes was further examined (Fig. 6C, 6D). The results demonstrated that miR-1 mimic decreased the phagocytic activity of hemocytes whereas miR-1 inhibitor increased it, indicating that miR-1 could negatively regulate the phagocytosis of hemocytes.

Furthermore, the role of miR-1 in antiviral and antibacterial responses was investigated. Compared with the control, miR-1 inhibitor significantly decreased the mortality of V. parahaemolyticus–infected shrimp (Fig. 7A) and reduced the copies of V. parahaemolyticus in tissues (Fig. 7C). Consistent with this, miR-1 mimic significantly elevated the bacterial load in tissues (Fig. 7C), and the final cumulative mortality of miR-1 mimic–injected shrimp infected by V. parahaemolyticus was higher than that of the control, although difference between the mortality curves did not reach significance (Fig. 7A). These suggested that miR-1 could facilitate bacterial infection in shrimp. In contrast, the mortality of WSSV-infected shrimp and the viral load of WSSV in tissues were significantly decreased after injection of miR-1 mimic and increased after injection of miR-1 inhibitor (Fig. 7B, 7D), suggesting that miR-1 could play a positive role in antiviral responses.

In this study, the crustacean counterpart of the miR-1 molecule was identified, and its putative precursor was verified. The target gene of miR-1 was determined as MyD88, an essential component of the Dorsal pathway (24, 44). MyD88 is the first intracellular adaptor that transmits TLR signals through interacting with the intracellular TIR domain of TLR (45, 46). Occurring at the starting point of an intracellular signal amplification cascade in the NF-κB pathway, changes in the activity of MyD88 due to changes in its expression can also be transmitted as signals (47, 48), which are expected to be efficiently amplified by downstream signal transduction molecules. As miR-1 significantly decreases the expression of MyD88, it is not surprising that miR-1 effectively inhibits activation of Dorsal. Moreover, the promoter of miR-1 contains a putative STAT binding site and can be directly bound and regulated by STAT. As expected, the JAK-STAT pathway upstream component JAK and inhibitor SOCS2 can also significantly affect the production of miR-1, confirming that miR-1 is a target gene of the JAK-STAT pathway. Therefore, the current study established a novel mechanism for the JAK-STAT pathway to regulate the Dorsal pathway in an invertebrate. Activation of the JAK-STAT pathway induces expression of miR-1, which further attenuates the activation of the Dorsal pathway through targeting MyD88. This could be a novel mechanism for balancing activities of the two signaling pathways in invertebrates.

As intracellular information transmitting and processing channels, signaling pathways are controlled by intrapathway feedback regulatory loops and interpathway cross-talks, which gives rise to sophisticated regulatory mechanisms underlying various life events (49). In mammals, several cross-talks between NF-κB and JAK-STAT signaling pathways have been established in the context of direct physical interaction at molecule level and functional convergence at promoter level (50). For instance, the unphosphorylated form of STAT-3 binds NF-κB and mediates its nuclear translocation to activate NF-κB target genes (51), whereas STAT-6 antagonizes transcriptional activation of NF-κB through preventing NF-κB binding at overlapping sites (52). More common ways of cross-talk depend on the mediation of several molecules that are involved in regulation of NF-κB or STAT activation, the expression of which are regulated by the other side of the liaison. These mediators include cytokines such as IL-1β, IL-6, IFNs, and TNF-α (53–56) and cytosolic signaling components such as SOCS, PKR, and OAS1A (57–60). Several miRNAs were also known to be involved in these processes. The case in point is that the STAT-3 target gene miR-146b inhibits the NF-κB–dependent production of IL-6, an activator of STAT-3, and subsequently inhibits the STAT-3 activation, constructing a negative feedback regulatory loop (61). These cross-talk mechanisms may contribute to coordinating the activation of the two signaling pathways in various biological processes in mammals. Compared with that in mammals, little is known about the interaction between NF-κB and JAK-STAT pathways in invertebrates. The current study may provide a novel viewpoint for understanding the complex regulatory network underlying biological processes in invertebrates. That is, some conserved miRNAs could be effective molecules mediating cross-talk between the two canonical signaling pathways at least in crustaceans.

Both JAK-STAT and NF-κB pathways are conserved in evolution and have been associated with certain aspects of the innate immune system. In Drosophila, the JAK-STAT pathway is required for antiviral immunity (62, 63), whereas the TLR-Dorsal pathway is mainly involved in immune responses against Gram-positive bacteria (64, 65). In shrimp, numerous studies have established the roles of the TLR-Dorsal pathway in immunity against both viruses and Gram-negative bacteria (20, 66). Sharing the same critical signaling components as those in mammals, mosquitoes, and Drosophila, the JAK-STAT pathway in shrimp is essentially implicated in antiviral responses, constructing the downstream part of the IRF/Vago/JAK-STAT–regulatory axis that may be similar to the mammalian IFN system (67, 68). Recent studies have suggested that the JAK-STAT pathway also plays important roles in antibacterial immunity (28). These indicated that the TLR-Dorsal and JAK-STAT pathways may overlappingly participate in some aspects of immune responses in shrimp. The current study informs us that activation of JAK-STAT pathway has an inhibitory effect on that of TLR-Dorsal pathway through a miR-1–mediated liaison, which could be a part of the regulatory network for coordinating the activation of the two pathways during immune responses.

Maintenance of immune homeostasis is essential for establishment of the defense mechanism and determination of the immune response pattern and involves the coordination of multiple intracellular signaling pathways. The abnormal shift of the balance of immune homeostasis or overactivation of a specific aspect of immunity may lead to autoimmune diseases and changes in susceptibility of the host to infection by a specific type of pathogen (69, 70). In mice, feedback interactions between NF-κB and STAT pathways play important roles in determination of immune response patterns (e.g., switch-like response to poly(I:C) or pulse-like response to LPS) (60). In the current model, when activated, the JAK-STAT pathway inhibits activation of TLR-Dorsal pathway through inducing miR-1, suggesting that, as a nodal molecule of the cross-talk, miR-1 may be implicated in maintenance of immune homeostasis. We demonstrated that miR-1 exerted significant effects on the expression of a set of immune related genes in vivo and on the defense of shrimp against viral and bacterial infections, indicating the important role of the cross-talk between TLR-Dorsal and JAK-STAT pathways in immune responses. Interestingly, in vivo experiments showed that miR-1 facilitated the infection by V. parahaemolyticus but inhibited the infection by WSSV, suggesting that miR-1 could play opposite roles in antimicrobial and antiviral responses. Considering that miR-1 is a negative regulator for the TLR-Dorsal pathway, the current results implied that in the context of immune homeostasis, shift of the balance to the enhancement of TLR-Dorsal pathway could enhance the immune response against bacteria but reduce that against virus, which needs further support. Taken together, the current study suggests a miR-1–mediated inhibitory effect of the JAK-STAT pathway on the NF-κB pathway, which could contribute to immune homeostasis and shaping immune responses in invertebrates. These findings may further our understanding of the invertebrate immune regulatory network and provide a basis for further exploration of cellular signaling system evolution.

The authors have no financial conflicts of interest.